Image processing method and system

Abstract

The present application provides an image processing method and system. A target image is acquired, wherein the target image includes a plurality of elements corresponding to pixels or voxels. The target image can be decomposed into at least one layer, wherein the at least one layer can include a low-frequency sub-image and a high-frequency sub-image. The at least one layer is transformed. The transformed layer can be reconstructed into a composite image.

Description

[0001] Description of the case

[0002] This application is a divisional application filed for the Chinese application with application date of December 29, 2016, application number 201680083009.0, and invention name “An Image Processing Method and System”.

[0003] Cross-references

[0004] This application claims priority to Chinese Application No. 201511027401.7, filed on December 31, 2015, Chinese Application No. 201511027173.3, filed on December 31, 2015, Chinese Application No. 201610510790.7, filed on July 1, 2016, and Chinese Application No. 201610584749.4, filed on July 22, 2016. The contents of the above applications are incorporated herein by reference. Technical Field

[0005] The present application relates to the field of image processing, and in particular to a method and system for enhancing images. Background Art

[0006] Imaging systems play an important role in the medical field. Imaging systems can generate and / or process medical images (e.g., CT images, PET images, MRI images, etc.) for medical diagnosis or radiotherapy. For example, a CT image of the breast can be used to screen for masses in the breast. Typically, medical images can be adjusted to facilitate doctors in identifying potential lesions. For example, images can be denoised and / or enhanced using different image processing techniques. However, adjustments to images may be inefficient and / or ineffective. For example, the edges of a region of interest may be missed; the grayscale values in the image may be uneven; or imaging noise may be enhanced. Therefore, image processing techniques that can enhance the contrast of an image and / or denoise an image are desirable for imaging systems. Summary of the Invention

[0007] One aspect of the present application relates to an image processing method. The method can be implemented on at least one machine, each machine having at least one processor and memory. The method can include one or more of the following operations: Acquire a target image, the target image comprising a plurality of elements, each corresponding to pixels or voxels; Decompose the target image into at least one layer, wherein the at least one layer comprises a low-frequency sub-image and a high-frequency sub-image; Transform the at least one layer; and Reconstruct the transformed layer into a composite image.

[0008] Another aspect of the present application relates to a non-transitory computer-readable medium comprising executable instructions that, when executed by at least one processor, cause the at least one processor to implement a method of image processing.

[0009] Another aspect of the present application relates to an image processing system. The system includes at least one processor and instructions. When executed by the at least one processor, the instructions cause the at least one processor to perform an image processing method. The system also includes a non-transitory computer-readable medium containing the instructions.

[0010] In some embodiments, acquiring the target image includes one or more of the following operations: acquiring an initial image; extracting a region of interest (ROI) based on the initial image; extracting an edge of the ROI based on the initial image; and determining an image of the ROI as the target image based on the ROI and the edge of the ROI.

[0011] In some embodiments, the region of interest is a breast, the region of interest edge is a breast edge, and the target image is a breast image.

[0012] In some embodiments, extracting the edge of the region of interest includes one or more of the following operations: denoising the initial image; preprocessing the denoised initial image based on a gradient transform; and detecting the edge of the region of interest.

[0013] In some embodiments, detecting the edge of the region of interest includes one or more of the following operations: The edge of the region of interest may be detected based on an OTSU algorithm or an iterative algorithm.

[0014] In some embodiments, extracting the region of interest includes one or more of the following operations: Segmenting the region of interest based on the OTSU algorithm or the watershed algorithm.

[0015] In some embodiments, the method further comprises one or more of the following operations: transforming the initial image into a logarithmic domain image.

[0016] In some embodiments, the low-frequency sub-image includes a preset region, the preset region includes multiple grayscale values, and the transforming of the layer includes one or more of the following operations: transforming the multiple grayscale values of the preset region.

[0017] In some embodiments, transforming multiple gray values of a preset region includes one or more of the following operations. Determine a reference edge in the low-frequency sub-image. Determine a characteristic curve based on the low-frequency sub-image. The characteristic curve represents the relationship between the distance and the gray value corresponding to the distance, where the distance refers to the distance between a first element in the low-frequency sub-image and a second element in the reference edge, and the first element corresponds to the second element. The gray value can be determined based on multiple gray values. Determine a transformation curve based on the characteristic curve, where the transformation curve represents the relationship between the gray value before transformation and the gray value after transformation. Update the multiple gray values of the preset region based on the transformation curve.

[0018] In some embodiments, determining the transformation curve includes one or more of the following operations. Divide the characteristic curve into N characteristic curve segments. Determine N transformation curve segments based on the N characteristic curve segments, where one characteristic curve segment corresponds to one transformation curve segment. Generate a transformation curve based on the N transformation curve segments.

[0019] In some embodiments, the determination of the N transformation curve segments includes one or more of the following operations. For the x-th transformation curve segment of the N transformation curve segments, calculate the slope of the x-th transformation curve segment based on the gray value of a preset point in the characteristic curve, the gray value of the initial point of the x-th characteristic curve, and the gray value of the end point of the x-th characteristic curve segment. The x-th characteristic curve segment corresponds to the x-th transformation curve segment, where x is an integer and 1 ≤ x ≤ N. Determining the gray value of the initial point in the x-th transformation curve segment includes one or more of the following operations. If x = 1, specify the gray value of the initial point in the x-th characteristic curve segment as the gray value of the initial point in the x-th transformation curve segment. If 1 < x ≤ N, determine the gray value of the initial point in the x-th transformation curve segment based on the gray value of the initial point of the (x - 1)-th transformation curve segment and the gray value change amount of the (x - 1)-th characteristic curve segment.

[0020] In some embodiments, the determination of the transformation curve further includes one or more of the following operations. Determine the gray value interval of the characteristic curve, in which at least one gray value is to be transformed, and the gray value interval corresponds to a part of the characteristic curve. Specify the maximum value or the minimum value of the gray value interval as the gray value of the preset point in the characteristic curve.

[0021] In some embodiments, the decomposition of the target image includes one or more of the following operations. Based on the first decomposition, decompose the target image into L layers, and each layer of the L layers includes a low-frequency sub-image and a high-frequency sub-image, where L ≥ 1. Based on the second decomposition, decompose the target image into L'+N image layers, and each layer of the L'+N layers includes a low-frequency sub-image and a high-frequency sub-image, where L' ≥ 1 and N ≥ 1.

[0022] In some embodiments, L is equal to L'.

[0023] In some embodiments, the first decomposition is a Laplace transform and the second decomposition is a wavelet transform.

[0024] In some embodiments, reconstruction of the transformed layer includes one or more of the following operations: Updating the low-frequency sub-image of the Lth layer generated from the first decomposition based on the low-frequency sub-image of the L'th layer generated from the second decomposition; Reconstructing a composite image based on the high-frequency sub-image of the Lth layer generated from the first decomposition and the updated low-frequency sub-image of the Lth layer.

[0025] In some embodiments, the method further comprises one or more of the following operations: enhancing the high frequency sub-image of the L layer produced by the first decomposition.

[0026] In some embodiments, reconstruction of the composite image comprises one or more of the following operations. For each of a plurality of iterations, upsampling the low-frequency sub-image of the (Li)th layer. For each of a plurality of iterations, updating the low-frequency sub-image of the (Li-1)th layer based on the upsampled low-frequency sub-image of the (Li)th layer and the high-frequency sub-image of the (Li)th layer, where 0≤i≤L-1. For each of a plurality of iterations, reconstructing the composite image based on the updated low-frequency sub-image of the first layer and the high-frequency sub-image of the first layer.

[0027] In some embodiments, upsampling the low-frequency sub-image of the (Li)th layer may include one or more of the following operations: Upsampling the low-frequency sub-image of the (Li)th layer based on bilinear interpolation or cubic interpolation.

[0028] In some embodiments, the method further includes one or more of the following operations: Updating the low-frequency sub-image of the L'th layer generated by the second decomposition based on the low-frequency sub-image of the (L'+N)th layer generated by the second decomposition and the high-frequency sub-images of the (L'+1)th to (L'+N)th layers generated by the second decomposition.

[0029] In some embodiments, the high-frequency sub-image includes a plurality of elements, and the transforming of the layer includes one or more of the following operations: generating a weight map for the high-frequency sub-image, wherein the weight map includes a plurality of weights corresponding to the plurality of elements; and updating the high-frequency sub-image based on the weight map.

[0030] In some embodiments, the high-frequency sub-image includes first-category elements and second-category elements, and the generation of the weight map includes one or more of the following operations. Determine the grayscale interval of the first-category elements in the high-frequency sub-image. Based on the grayscale interval of the first-category elements, determine the grayscale interval of the second-category elements in the high-frequency sub-image; map the grayscale interval of the first-category elements to [0,1]. Determine the weight of the first-category elements based on the mapped grayscale interval of the first-category elements. Map the grayscale interval of the second-category elements to (1, G], where G is a preset value. Determine the weight of the second-category elements based on the mapped grayscale interval of the second-category elements. Generate a weight map based on the weight of the first-category elements and the weight of the second-category elements.

[0031] In some embodiments, determining the grayscale interval of the first-category element includes one or more of the following operations: determining an initial grayscale interval of the first-category element based on a grayscale threshold; modifying the initial grayscale interval of the first-category element; and adjusting the initial grayscale interval of the first-category element based on the modified grayscale interval of the first-category element.

[0032] In some embodiments, adjusting the initial grayscale interval of the first type of elements includes one or more of the following operations: calculating a first threshold based on the modified grayscale interval of the first type of elements; determining the grayscale interval of the first type of elements to be [0, first threshold].

[0033] In some embodiments, the transformation of the high-frequency sub-image includes one or more of the following operations: Multiplying the grayscale value of the high-frequency sub-image by the grayscale value of the weight map.

[0034] In some embodiments, the transformation of the layer includes one or more of the following operations: Transforming the high frequency sub-image by linear / non-linear enhancement or denoising.

[0035] In some embodiments, the transformation of the layer comprises one or more of the following operations: Transforming the low frequency sub-image by linear / non-linear enhancement or denoising. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The present application is further described based on exemplary embodiments. These exemplary embodiments are described in detail with reference to the accompanying drawings. These embodiments are non-limiting exemplary embodiments, in which like reference numerals represent similar structures in the several views of the drawings, wherein:

[0037] Figure 1 is a schematic diagram of an exemplary imaging system according to some embodiments of the present application;

[0038] Figure 2-A is a schematic diagram of an exemplary image processing system according to some embodiments of the present application;

[0039] Figure 2-Bis a flowchart of an exemplary process of processing an image according to some embodiments of the present application;

[0040] Figure 3-A is a schematic diagram of an exemplary image acquisition module according to some embodiments of the present application;

[0041] Figure 3-B is a flowchart of an exemplary process of acquiring an image according to some embodiments of the present application;

[0042] Figure 4-A is a flowchart of an exemplary process of decomposing an image according to some embodiments of the present application;

[0043] Figure 4-B is a schematic diagram of an exemplary L layer decomposed by a first decomposition unit according to some embodiments of the present application;

[0044] Figure 4-C is a schematic diagram of an exemplary L'+N layer decomposed by a second decomposition unit according to some embodiments of the present application;

[0045] Figure 5-A is a schematic diagram of an exemplary grayscale value conversion unit according to some embodiments of the present application;

[0046] Figure 5-B is a flowchart of an exemplary process of transforming an image according to some embodiments of the present application;

[0047] Figure 6 is a schematic diagram of exemplary characteristic curves according to some embodiments of the present application;

[0048] Figure 7 is a schematic diagram of an exemplary characteristic curve divided into a plurality of characteristic curve segments according to some embodiments of the present application;

[0049] Figure 8 is a schematic diagram of exemplary transformation curves according to some embodiments of the present application;

[0050] Figure 9-A is a schematic diagram of an exemplary weight transformation unit according to some embodiments of the present application;

[0051] Figure 9-B is a flowchart of an exemplary process of transforming a target image based on a weight map according to some embodiments of the present application;

[0052] Figure 9-C is a flowchart of an exemplary process of determining a weight map according to some embodiments of the present application;

[0053] Figure 10is a schematic diagram illustrating an exemplary process of generating a weight graph;

[0054] Figure 11-A is a flowchart of an exemplary process of layer-based reconstructing a synthetic image according to some embodiments of the present application;

[0055] Figure 11-B is a flowchart of an exemplary process of reconstructing a low-frequency sub-image of the L'th layer generated by the second decomposition according to some embodiments of the present application;

[0056] Figure 11-C is a flowchart of an exemplary process of reconstructing a composite image based on L layers generated by the first decomposition according to some embodiments of the present application. DETAILED DESCRIPTION

[0057] In the following detailed description, many specific details are set forth by way of example in order to provide a thorough understanding of the relevant disclosure. However, it should be understood by those skilled in the art that the present application can be implemented without these details. In other cases, in order to avoid unnecessarily obscuring some aspects of the present disclosure, the present disclosure describes well-known methods, procedures, systems, components and / or circuits at a relatively high level and without adding details. It will be apparent to those skilled in the art that various changes can be made to the disclosed embodiments. In addition, without departing from the principle and scope of the present application, the general principles defined in the present application may be applied to other embodiments and application scenarios. Therefore, the present application is not limited to the embodiments shown, but is the widest scope consistent with the claims.

[0058] It should be understood that the terms "system," "engine," "module," and / or "unit" used herein are a method for distinguishing different components, elements, parts, portions, or assemblies at different levels. However, other expressions that can achieve the same purpose may be used to replace the above terms.

[0059] It should be understood that when a device, unit, or module is described as being "on," or "connected to," or "coupled to" another device, unit, or module, unless the context clearly indicates otherwise, these devices, units, or modules may be directly connected to or coupled to the other device, unit, or module, or be communicatively connected to the other device, unit, or module, or other intermediate devices, units, or modules may exist. In this application, the term "and / or" may include any one or more of the relevant listed items or any combination thereof.

[0060] The terms used in this application are only for describing specific embodiments and do not limit the scope of this application. Unless the context clearly indicates an exception, the words "one", "a", "an" and "the" used in this application do not specifically refer to the singular and may also include the plural. It should also be understood that the terms "including" and / or "comprising" in this application only detail the existence of a whole, device, behavior, stated features, steps, elements, operations and / or parts, but do not exclude the existence or addition of one or more other wholes, devices, behaviors, features, steps, elements, operations, parts and / or their combinations.

[0061] The following description is provided for illustrative purposes to facilitate a better understanding of the methods or systems for image processing or enhancement. It should be noted that "image" may refer to a medical image, a still picture, or a video frame. It should be understood that this is not intended to limit the scope of this application. Those skilled in the art may make certain variations, modifications, and / or alterations based on the guidance of this application. Such variations, modifications, and / or alterations do not depart from the scope of this application.

[0062] Figure 1 1 is a block diagram of an exemplary imaging system according to some embodiments of the present application. The imaging system can generate images of an object. As shown in the figure, the imaging system includes an imaging device 110, a controller 120, an image processing system 130, a memory 140, and an input / output device 150.

[0063] The imaging device 110 can scan an object and generate a plurality of data related to the object. In some embodiments, the imaging device 110 can be a medical imaging device, such as a PET device, a SPECT device, a CT device, an MRI device, an X-ray device (e.g., a full-field digital mammography device (FFDM)), a digital breast tomosynthesis (DBT) device, a digital subtraction angiography (DSA) system, a magnetic resonance angiography (MRA) system, a computed tomography angiography (CTA), a digital radiography (DR) system, etc., or any combination thereof (e.g., a PET-CT device, a PET-MRI device, or a SPECT-MRI device).

[0064] In some embodiments, the imaging device 110 may include a scanner to scan an object and obtain information related to the object. In some embodiments, the imaging device 110 may be a radioactive scanning device. The radioactive scanning device may include a radiation source to emit radiation toward the scanned object. The radiation may include, for example, particle radiation, photon radiation, or any combination thereof. The particle radiation may include neutrons, protons, electrons, alpha rays, muons, heavy ions, or any combination thereof. The photon radiation may include X-rays, gamma rays, ultraviolet rays, lasers, or any combination thereof.

[0065] In some embodiments, the photon rays may be X-rays, and the imaging device 110 may be a CT system, a digital radiography (DR) system, a multimodal system, or the like, or any combination thereof. Exemplary multimodal systems may include a CT-PET system, a SPECT-MRI system, or the like. In some embodiments, the imaging device 110 may include an X-ray generating unit (not shown) and an X-ray detecting unit (not shown). In some embodiments, the imaging device 110 may include a photon detector to capture photons generated from the scanned object. In some embodiments, the photon detector may include a scintillator and / or a photodetector, and the imaging device 110 may be a PET system, or a multimodal system (e.g., a PET-CT system, a PET-MRI system, etc.). In some embodiments, the imaging device 110 may include a main magnetic field generator, a plurality of gradient coils, a radio frequency (RF) transmitter, and / or an RF receiver. The imaging device 110 may be an MRI system, or a multimodal system (e.g., a PET-MRI system, a SPECT-MRI system, etc.).

[0066] The controller 120 can control the imaging device 110, the image processing system 130, the memory 140, and / or the input / output device 150. The controller 120 can also control communications between the imaging device 110, the image processing system 130, the memory 140, the input / output device 150, and / or the network 160. The controller 120 can receive information from or send information to the imaging device 110, the memory 140, the input / output device 150, and / or the image processing system 130. For example, the controller 120 can receive commands from the input / output device 150 provided by a user. The controller 130 can process the information input by the user via the input / output device 150 and convert the information into one or more commands. For another example, the controller 120 can control the imaging device 110, the input / output device 150, and / or the image processing system 130 based on the received commands or the converted commands. For another example, the controller 120 can receive image signals or data related to a subject from the imaging device 110. For another example, the controller 120 can send the image signals or data to the image processing system 130. For another example, the controller 120 may receive processed data or a constructed image from the image processing system 130. For another example, the controller 120 may send the processed data or the constructed image to the input / output device 150 for display. For another example, the controller 120 may send the processed data or the constructed image to the memory 140 for storage. For another example, the controller 120 may read information from the memory 140 and send the information to the image processing system 130 for processing. In some embodiments, the controller 120 may include a computer, a program, an algorithm, software, a storage device, one or more interfaces, etc. Exemplary interfaces may include interfaces to the imaging device 110, the input / output device 150, the image processing system 130, the memory 140, and / or other modules or units in the imaging system.

[0067] In some embodiments, the controller 120 can receive commands provided by a user, including, for example, an imaging technician, a physician, etc. Exemplary commands can relate to the duration of a scan, the position of the subject, the position of the bed on which the subject rests, the operating conditions of the imaging device 110, specific parameters that can be used in image processing, etc., or any combination thereof. In some embodiments, the controller 120 can control the image processing system 130 to select different algorithms for processing the images.

[0068] The image processing system 130 can process information received from the imaging device 110, the controller 120, the memory 140, the network 160, and / or the input / output device 150. In some embodiments, the image processing system 130 can generate one or more images based on the information. In some embodiments, the image processing system 130 can process one or more images. The images processed by the image processing system 130 can include 2D images and / or 3D images. The image processing system 130 can transmit the images to the input / output device 150 for display or to the memory 140 for storage. In some embodiments, the image processing system 130 can perform operations including, for example, image preprocessing, image reconstruction, image enhancement, image correction, image synthesis, lookup table creation, etc., or any combination thereof. In some embodiments, the image processing system 130 can process data based on algorithms including, for example, filtered backprojection algorithms, fan-beam reconstruction, iterative reconstruction, grayscale value transformation, filtering, wavelet transforms, Laplace transforms, etc., or any combination thereof.

[0069] In some embodiments, the image processing system 130 may include one or more processors to perform the processing operations disclosed herein. The processor may include a microcontroller, a microprocessor, a reduced instruction set computer (RISC), an application-specific integrated circuit (ASIC), an application-specific instruction set processor (ASIP), a central processing unit (CPU), a graphics processing unit (GPU), a physical processing unit (PPU), a microcontroller unit, a digital signal processor (DSP), a field programmable gate array (FPGA), an advanced RISC machine (ARM), a programmable logic device, and any circuit and processor capable of performing one or more functions, or any combination thereof.

[0070] In some embodiments, the image processing system 130 may further include a memory configured to store data and / or instructions. In some embodiments, the memory may include mass storage, removable storage, volatile read-write memory, read-only memory (ROM), or any combination thereof. Exemplary mass storage may include a magnetic disk, an optical disk, a solid-state disk, or the like. Exemplary removable storage may include a flash drive, a floppy disk, an optical disk, a memory card, a compact disk, a magnetic tape, or the like. Exemplary volatile read-write memory may include random access memory (RAM). Exemplary RAM may include dynamic RAM (DRAM), double data rate synchronous dynamic RAM (DDR SDRAM), static RAM (SRAM), thyristor RAM (T-RAM), and zero-capacitance RAM (Z-RAM). Exemplary ROM may include masked ROM (MROM), programmable ROM (PROM), erasable programmable ROM (EPROM), electronically erasable programmable ROM (EEPROM), compact disc ROM (CD-ROM), and digital versatile disk ROM. In some embodiments, the memory may store one or more programs and / or instructions that are executed by the processor of the image processing system 130 to implement the exemplary methods described herein. For example, the memory may store programs and / or instructions executed by the processor of the image processing system 130 to decompose an image, transform the decomposed image, and / or reconstruct an image. For example, the ROM may store a decomposition algorithm (e.g., a Laplace algorithm) for the image processing system 130 to decompose an image.

[0071] In some embodiments, image data regarding a region of interest (ROI) may be processed by the image processing system 130. In some embodiments, the image processing system 130 may improve image quality, enhance image contrast, reduce or remove image artifacts, reduce or eliminate image noise, and / or enhance ROI edge information. Image artifacts may include oscillation artifacts, speckle artifacts, or any combination thereof. ROI edge information may refer to information regarding the edge of the ROI (e.g., grayscale value, contrast, brightness, etc.), where the edge of the ROI may include, for example, a breast edge, a tumor edge, or any combination thereof.

[0072] In some embodiments, the image processing system 130 may generate control signals related to the configuration of the imaging device 110. In some embodiments, the results generated by the image processing system 130 may be provided to other modules or units in the system, including, for example, a database (not shown) and a terminal (not shown), via the network 160. In some embodiments, data from the image processing system 130 may be sent to the memory 140 for storage.

[0073] Memory 140 can store information sent from imaging device 110, controller 120, image processing system 130, input / output device 150, and / or external data storage via network 160. The stored information can include numerical values, signals, images, object information, instructions, algorithms, etc., or a combination thereof. Memory 140 can refer to system memory (e.g., a disk) that is provided integrally (i.e., substantially non-removable), or memory that is removably connected to the system via, for example, a port (e.g., a USB port, a FireWire port, etc.). Memory 140 can include, for example, a hard disk, a floppy disk, electronic memory, random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), bubble memory, thin film memory, magnetic plated wire memory, phase change memory, flash memory, cloud storage, etc., or a combination thereof. Memory 140 can be connected to or communicate with one or more components of the imaging system. In some embodiments, memory 140 can be operatively connected to one or more virtual storage resources (e.g., cloud storage, a virtual private network, other virtual storage resources, etc.) via network 160.

[0074] The input / output device 150 can receive or output information. In some embodiments, the input / output device 150 may include a terminal, a keyboard, a touch screen, a cursor control device, a remote controller, or the like, or any combination thereof. The terminal may include, for example, a control panel, a mobile device (e.g., a smart phone, a tablet computer, a laptop computer, etc.), a personal computer, other devices, or the like, or any combination thereof. Other devices may include devices that can work independently, or processing units or processing modules assembled in another device. The cursor control device may include a mouse, a trackball, or cursor direction keys to transmit direction information and command selections to, for example, the image processing system 130 and control cursor movement on a display device.

[0075] The input and / or output information may include programs, software, algorithms, data, text, numbers, images, voice, etc., or any combination thereof. For example, a user may input some initial parameters or conditions to start the imaging process. For another example, some information may be imported from external resources, and the external resources may include, for example, a floppy disk, a hard disk, a wired terminal, a wireless terminal, etc., or any combination thereof. In some embodiments, the input and / or output information may also include alphanumeric and / or other key information, which may be input via a keyboard, a touch screen (e.g., with tactile or haptic feedback), voice input, image input, eye tracking input, a brain monitoring system, or any other similar input mechanism. The output information may be sent to a memory 140, a display (not shown), a printer (not shown), a computing device, etc., or a combination thereof.

[0076] In some embodiments, the input / output device 150 may include a user interface. The user interface may be a user interaction interface, a graphical user interface (GUI), a user-defined interface, or the like. The graphical user interface may allow a user to interact with other components (e.g., the imaging device 110 and / or the controller 120). For example, the graphical user interface may facilitate user input of parameters and intervention in the image processing process. In some embodiments, the input / output device 150 may include a display. The display may include a liquid crystal display (LCD), a light emitting diode (LED)-based display, a flat panel display or a curved screen (or television), a cathode ray tube (CRT), a 3D display, a plasma display panel, or the like, or any combination thereof.

[0077] The network 160 may facilitate communications between the imaging device 110, the controller 120, the image processing system 130, the memory 140, and / or the input / output device 150. For example, information may be transmitted from the imaging device 110 to the image processing system 130 via the network 160. As another example, information processed and / or generated by the image processing system 130 may be transmitted to the memory 140 and / or the input / output device 150 via the network 160.

[0078] In some embodiments, the network 160 may be a wired network, a nanoscale network, a near field communication (NFC), a body area network (BAN), a personal area network (PAN, e.g., Bluetooth, Z-wave, wireless personal area network, wireless USB), a near area network (NAN), a local wireless network, a backbone network, a metropolitan area network (MAN), a wide area network (WAN), an internet area network (IAN or cloud), or the like, or any combination thereof. The present application may also utilize known communication technologies that provide a medium for transmitting data between separate devices. In some embodiments, the network 160 may be a single network or a combination of various networks. The network 160 may include, but is not limited to, a local area network, a wide area network, a public network, a private network, a wireless local area network, a virtual network, a metropolitan area network, a public switched telephone network, or a combination thereof. In some embodiments, the network 160 may include various network access points, such as wired or wireless access points, base stations, or network switching points, through which data sources may connect to the network 160 and information may be sent via the network 160.

[0079] In some embodiments, two or more of the imaging device 110, controller 120, image processing system 130, memory 140, and input / output device 150 may be directly connected or in communication with one another. In some embodiments, the imaging device 110, controller 120, image processing system 130, memory 140, and input / output device 150 may be connected or in communication with one another via a network 160. In some embodiments, the imaging device 110, controller 120, image processing system 130, memory 140, and input / output device 150 may be connected or in communication with one another via an intermediate unit (not shown). The intermediate unit may be a visible component or an invisible field (radio, optical, acoustic, electromagnetic induction, etc.). The connection between the different units may be wired or wireless. Wired connections may include metal cables, optical cables, hybrid cables, interfaces, etc., or any combination thereof. Wireless connections may include local area networks (LANs), wide area networks (WANs), Bluetooth, wireless personal area networks, near field communications (NFC), etc., or any combination thereof. Network 160 may be used in conjunction with the systems described herein and is not intended to be exhaustive or limiting.

[0080] The imaging system described here is for illustrative purposes only and is not intended to limit the scope of this application. Applications of the imaging system can be found in different fields, such as medicine or industry. The imaging device 110 can be used for internal inspection of components, including, for example, flaw detection, safety scanning, fault analysis, metrology, assembly analysis, gap analysis, wall thickness analysis, etc., or any combination thereof. For those skilled in the art, after understanding the basic principles of the connection between different components, the components and connections between units can be modified or changed without departing from the principles. These modifications and changes are still within the scope of the above-mentioned current application. In some embodiments, these components can be independent, and in some embodiments, parts of these components can be integrated into one component to work together.

[0081] Figure 2-A FIG. 1 is a schematic diagram of an exemplary image processing system 130 according to some embodiments of the present application. Figure 2-A As shown, the image processing system 130 may include an image acquisition module 201 , a decomposition module 202 , a transformation module 203 and a reconstruction module 204 .

[0082] In some embodiments, the image acquisition module 201 can acquire an image. An image can be a 2D image or a 3D image. A 2D image may include multiple pixels. A 3D image may include multiple voxels. A pixel / voxel may have a corresponding value, including, for example, brightness, color, grayscale, or any combination thereof. For simplicity, a pixel or voxel may be referred to as an element. An element may refer to a pixel of a 2D image or a voxel of a 3D image. In some embodiments, the image acquisition module 201 can obtain a signal and / or data representing the image. The signal can take any of a variety of forms, including electromagnetic, optical, or any suitable combination thereof. In some embodiments, the image data may include raw data, processed data, control data, interaction data, image data, video data, analog data, digital data, or any combination thereof. In some embodiments, the acquired image may include an initial image, a region of interest (ROI) of the initial image, any image generated during the image processing process, or any combination thereof.

[0083] The initial image may refer to an image initially acquired by the image acquisition module 201. In some embodiments, the initial image may be generated by the imaging device 110. For example, the initial image may be a raw image captured by a CT system, a PET system, an MRI system, or the like, or any combination thereof. In some embodiments, the initial image may be obtained from the memory 140, the input / output device 150, or an external data storage device via the network 160. For example, the initial image may be a processed image pre-stored in the memory 140. In some embodiments, the initial image may be processed by the image acquisition module 201, and a target image may be generated. The target image may refer to an image decomposed, transformed, reconstructed, and / or enhanced by the image processing system 130.

[0084] The region of interest image may include a region of interest of the initial image. The region of interest may refer to a portion of the initial image, including information of interest in the initial image, including, for example, tissue of interest, organ of interest, background of interest, lesion of interest, any region of interest, etc., or any combination thereof. In some embodiments, the region of interest may be extracted from the initial image, and a region of interest image may be obtained. For example, an organ of interest may be extracted from the initial image, and an organ image may be generated. The organ image may include a lung, a breast, a portion of a bone, a portion of a muscle, an eye, any part of the body, etc., or any combination thereof. By way of example only, the organ image may be a breast image extracted from a chest CT image.

[0085] In some embodiments, the image acquired by the image acquisition module 201 may be obtained from the imaging device 110, the memory 140, the input / output device 150, an external data storage device via the network 160, any component of the image processing system 130, or any combination thereof. In some embodiments, the acquired image may be processed by the decomposition module 202, the transformation module 203, and / or the reconstruction module 204. For example, the acquired image may be decomposed by the decomposition module 202 to enhance the image. In some embodiments, the acquired image may be stored in the memory 140, displayed by the input / output device 150, transmitted to a terminal or an external data storage device via the network 160, or any combination thereof.

[0086] In some embodiments, the decomposition module 202 can decompose the target image acquired by the image acquisition module 201. Figure 2-A As shown, the decomposition module 202 may include a first decomposition unit 202-1, a second decomposition unit 202-2, and / or an Nth decomposition unit 202-N, where N is an integer. In some embodiments, at least two decomposition units may use the same decomposition algorithm. In some embodiments, at least two decomposition units may use different decomposition algorithms.

[0087] In some embodiments, the decomposition module 202 may decompose the image into one or more layers. In some embodiments, image decomposition may refer to dividing or decomposing the image into one or more image layers based on the grayscale values of the image elements, the frequency of the image, etc. An image layer may include two or more sub-images. For example, an image layer may include a low-frequency sub-image and a high-frequency sub-image. In some embodiments, the low-frequency sub-image and the high-frequency sub-image may be determined based on one or more frequency thresholds. For example, the sub-images with a frequency lower than or equal to the frequency threshold T may be selected. f The sub-image is determined as a low-frequency sub-image. For another example, the frequency is greater than or equal to the frequency threshold T f The sub-image of can be determined as a high-frequency sub-image. f It can be predetermined according to the default setting of the image processing system 130 or determined by the user through the GUI of the input / output device 150. In some embodiments, the threshold T can be adjusted based on the processing efficiency of the image in the image processing system 130. f In some embodiments, the decomposition module 202 may further decompose the sub-image into one or more layers. For example, the decomposition module 202 may decompose the low-frequency sub-image or the high-frequency sub-image into one or more layers. In some embodiments, the decomposition module 202 may decompose the low-frequency sub-image in the first layer into the low-frequency sub-image and the high-frequency sub-image in the second layer. In some embodiments, the decomposition module 202 may decompose the low-frequency sub-image in the second layer into the low-frequency sub-image and the high-frequency sub-image in the third layer.

[0088] In some embodiments, the first decomposition unit 202-1 may decompose an image (or sub-image) based on a first decomposition algorithm. In some embodiments, the second decomposition unit 202-2 may decompose an image (or sub-image) based on a second decomposition algorithm. In some embodiments, the Nth decomposition unit 202-N may decompose an image (or sub-image) based on an Nth decomposition algorithm. In some embodiments, the decomposition algorithm may include wavelet transform, bilateral filtering, Fourier transform, discrete cosine transform, Laplace transform, any algorithm capable of decomposing an image, or any combination thereof. Wavelet transform may include continuous wavelet transform, discrete wavelet transform (DWT), fast wavelet transform (FWT), lifting scheme and generalized lifting scheme, wavelet packet decomposition, static wavelet transform, fractional Fourier transform, fractional wavelet transform, or any combination thereof. In some embodiments, at least two decomposition units may use different decomposition algorithms. For example, the first decomposition unit 202-1 may decompose an image based on a Laplace transform, and the second decomposition unit 202-2 may decompose an image based on a wavelet transform. For another example, the first decomposition unit 202 - 1 may decompose the image based on wavelet transform, and the second decomposition unit 202 - 2 may decompose the image based on Laplace transform.

[0089] In some embodiments, the first decomposition unit 202-1 may decompose the image (or sub-image) into one or more layers based on a first decomposition algorithm. For example, the first decomposition unit 202-1 may decompose the image (or sub-image) into two or more sub-images in a first layer based on the first decomposition algorithm. In some embodiments, the first decomposition unit 202-1 may also decompose the sub-image in the first layer into two or more sub-images in a second layer based on the first decomposition algorithm. In some embodiments, the first decomposition unit 202-1 may also decompose the sub-image in the Lth layer into two or more sub-images in the (L+1)th layer based on the first decomposition algorithm, where L is an integer greater than 2. For example, the first decomposition module 202-1 may decompose the image (or sub-image) into three layers based on a wavelet transform, where each layer may include a high-frequency sub-image and a low-frequency sub-image, the sub-image in the second layer may be generated from the low-frequency sub-image in the first layer, and the sub-image in the third layer may be generated from the low-frequency sub-image in the second layer.

[0090] In some embodiments, the second decomposition unit 202-2 may decompose the image (or sub-image) into one or more layers based on a second decomposition algorithm. For example, the second decomposition unit 202-2 may decompose the image (or sub-image) into two or more sub-images in the first layer based on the second decomposition algorithm. In some embodiments, the second decomposition unit 202-2 may also decompose the sub-image in the first layer into two or more sub-images in the second layer based on the second decomposition algorithm. In some embodiments, the second decomposition unit 202-2 may also decompose the sub-image in the Lth layer into two or more sub-images in the (L+1)th layer based on the second decomposition algorithm, where L is an integer greater than 2. For example, the second decomposition unit 202-2 may decompose the image (or sub-image) into five layers based on a Laplace transform, where each layer may include a high-frequency sub-image and a low-frequency sub-image. The sub-image in the second layer may be generated from the low-frequency sub-image in the first layer. The sub-image in the third layer may be generated from the low-frequency sub-image in the second layer. The sub-image in the fourth layer may be generated from the low-frequency sub-image in the third layer. The sub-image in the fifth layer may be generated from the low-frequency sub-image in the fourth layer.

[0091] In some embodiments, the Nth decomposition unit 202-N may decompose the image (or sub-image) into one or more layers based on the Nth decomposition algorithm. For example, the Nth decomposition unit 202-N may decompose the image (or sub-image) into two or more sub-images in the first layer based on the Nth decomposition algorithm. In some embodiments, the Nth decomposition unit 202-N may also decompose the sub-image in the first layer into two or more sub-images in the second layer based on the Nth decomposition algorithm. In some embodiments, the Nth decomposition unit 202-N may also decompose the sub-image in the Lth layer into two or more sub-images in the (L+1)th layer based on the Nth decomposition algorithm, where L is an integer greater than 2.

[0092] In some embodiments, the image may be decomposed into one or more layers by the first decomposition unit 202-1. In some embodiments, the sub-images in one layer generated by the first decomposition unit 202-1 may be further decomposed into one or more layers including sub-images by the second decomposition unit 202-2. In some embodiments, the sub-images generated by the second decomposition unit 202-2 may be further decomposed by the Nth decomposition unit 202-N. For example, the first decomposition unit 202-1 may decompose the image into a first layer including low-frequency sub-images and high-frequency sub-images. The low-frequency sub-images generated by the first decomposition unit 202-1 may be further decomposed into low-frequency sub-images and high-frequency sub-images in a second layer by the second decomposition unit 202-2. The low-frequency sub-images in the second layer generated by the second decomposition unit 202-2 may be decomposed into low-frequency sub-images and high-frequency sub-images by the Nth decomposition unit 202-N. In some embodiments, the decomposition module 202 may decompose the image into at least two layers (e.g., five layers) based on the same decomposition algorithm or different decomposition algorithms.

[0093] In some embodiments, the image and / or sub-image generated by the decomposition module 202 may be acquired by the image acquisition module 201, processed by the transformation module 203, provided to the reconstruction module 204, stored in the memory 140, transmitted via the network 160, etc., or any combination thereof. For example, two or more sub-images may be reconstructed into a processed image by the reconstruction module 204.

[0094] The transformation module 203 can transform the target image or sub-image. The image (or sub-image) to be transformed can be obtained from the image acquisition module 201 or the decomposition module 202. In some embodiments, the transformation module 203 can change the value of one or more elements in the image (or sub-image). Figure 2-A As shown, the transformation module 203 may include a grayscale value transformation unit 203-1, a weight transformation unit 203-2, an enhancement unit 203-3 and an upsampling unit 203-4.

[0095] The grayscale value conversion unit 203-1 can convert the grayscale values of one or more elements in the target image (or sub-image). In some embodiments, the grayscale value conversion unit 203-1 can convert the grayscale values of one or more elements in the target image (or sub-image) to adjust the grayscale values of the elements in the region of interest, thereby improving the quality of the target image (or sub-image), reducing noise, etc., or any combination thereof. In some embodiments, the grayscale value conversion unit 203-1 can convert the image (or sub-image) based on one or more conversion techniques. The conversion techniques can be based on conversion functions, conversion rules, conversion curves, etc., or any combination thereof. Figure 5-A and Figure 8An exemplary conversion curve is shown. In some embodiments, the grayscale value conversion unit 203-1 can convert different grayscale values based on different conversion techniques. In some embodiments, the grayscale value can be represented by an integer. The grayscale value can be limited to a certain range. For example, the range of the grayscale value can be between 0 and 1, or between 1 and 255.

[0096] In some embodiments, a grayscale value may represent the intensity of an element in a grayscale image. In some embodiments, a grayscale value may represent the intensity of an element in a single color channel of a color image (or sub-image). The color image (or sub-image) may have a color space that includes a red / green / blue (RGB) space, a hue / saturation / luminance (HIS) space, a hue / saturation / luminance (HSL) space, a hue / saturation / value (HSV) space, an International Aviation Committee (CIE) space, a hue / saturation / intensity (HSI) space, or any other color space that can represent human color perception, or any combination thereof. For example, a grayscale value may represent the intensity in the red channel of a color image in an RGB space.

[0097] In some embodiments, the transformation technique may be modified based on the transformation result generated by the grayscale value transformation unit 203-1, stored in the memory 140, or obtained from the network 160. The transformation technique may be used to compress or enhance the grayscale value of one or more elements. In some embodiments, the transformation curve may be generated based on a characteristic curve. The characteristic curve may be determined based on the initial image, the region of interest, the edge of the region of interest, the region of interest image, or a sub-image. For example, the characteristic curve may be determined based on a low-frequency sub-image (e.g., a low-frequency sub-image generated by the decomposition module 202). In some embodiments, the characteristic curve may be represented by a minimum distance (s) in the horizontal axis and a corresponding grayscale value of an element in the low-frequency sub-image in the vertical axis. In some embodiments, the horizontal axis may refer to the X-axis of a Cartesian coordinate. In some embodiments, the vertical axis may refer to the Y-axis of a Cartesian coordinate.

[0098] The weight transformation unit 203-2 can transform the grayscale values of elements in the target image (or sub-image) based on one or more weights or weight maps. In some embodiments, the weight transformation unit 203-2 can transform the grayscale values to reduce noise, reduce speckle artifacts, improve image contrast, etc., or any combination thereof. In some embodiments, the weight map can be determined based on the target image (or sub-image). For example, the target sub-image can be a high-frequency sub-image generated by the decomposition module 202, and the weight map can be generated based on the values of one or more elements in the high-frequency sub-image (e.g., grayscale values, brightness values, etc.). For another example, the sub-image can be a low-frequency sub-image generated by the decomposition module 202, and the weight map can be generated based on the values of one or more elements in the low-frequency sub-image (e.g., grayscale values, brightness values, etc.).

[0099] The enhancement unit 203-3 can enhance the target image (or sub-image). In some embodiments, the enhancement unit 203-3 can enhance the image (or sub-image) based on linear enhancement and / or nonlinear enhancement. Linear enhancement can include maximum-minimum contrast technology, percentage contrast technology, segmented contrast technology, etc., or any combination thereof. Nonlinear enhancement can include histogram equalization technology, adaptive histogram equalization technology, homomorphic filtering technology, unsharp masking technology, etc., or any combination thereof. In some embodiments, the enhancement unit 203-3 can enhance the target sub-image generated by the decomposition module 202. For example, the enhancement unit 203-3 can enhance the high-frequency sub-image or the low-frequency sub-image decomposed by the first decomposition unit 202-1, the second decomposition unit 202-2, or the Nth decomposition unit 202-N.

[0100] Upsampling unit 203-4 can upsample the target image (or sub-image). In some embodiments, upsampling unit 203-4 can upsample the target image (or sub-image) based on one or more interpolation operations, such as piecewise constant interpolation, linear interpolation, polynomial interpolation, spinal interpolation, or any combination thereof. In some embodiments, upsampling unit 203-4 can upsample one or more sub-images. For example, high-frequency sub-images and / or low-frequency sub-images generated by decomposition module 202 can be interpolated by upsampling unit 203-4.

[0101] In some embodiments, the image and / or sub-image transformed by the transformation module 203 can be decomposed by the decomposition module 202, reconstructed by the reconstruction module 204, stored in the memory 140, displayed by the input / output device 150, or transmitted via the network 160. In some embodiments, two or more sub-images transformed by the same unit in the transformation module 203 can be reconstructed into a composite image. For example, two or more sub-images transformed by the grayscale value transformation unit 203-1 can be reconstructed into a composite image by the reconstruction module 204. In some embodiments, the image and / or sub-image transformed by two or more units in the transformation module 203 can be reconstructed. For example, a low-frequency sub-image transformed by the grayscale value transformation unit 203-1 and a high-frequency sub-image transformed by the weight transformation unit 203-2 can be reconstructed into a composite image.

[0102] The reconstruction module 204 may reconstruct an image based on two or more images (or sub-images). The images (or sub-images) may be obtained from the image acquisition module 201, the decomposition module 202, the transformation module 203, the memory 140, the input / output device 150, or an external data storage device via the network 160. In some embodiments, the reconstruction module 204 may reconstruct the image based on a technique including, for example, a filtered back projection algorithm, an iterative reconstruction algorithm, a local reconstruction algorithm, a multiple additive regression tree algorithm, a randomized transformation algorithm, a Fourier transform algorithm, or any combination thereof.

[0103] Figure 2-B FIG2 is a flow chart of an exemplary process of processing an image according to some embodiments of the present application. The process may include acquiring an image 211 , decomposing the image 212 , transforming the image 213 , and reconstructing the image 214 .

[0104] At 211, an image may be acquired by the image acquisition module 201. The image may be acquired from any component of the imaging device 110, the memory 140, the input / output device 150, the network 160, the image processing system 130, or any combination thereof. The acquired image may include an initial image, an image of a region of interest, any image generated during the image processing process, or any combination thereof. In some embodiments, the image acquisition module 201 may extract information from the initial image and generate a target image. For example, the image acquisition module 201 may extract a region of interest from the initial image and generate a region of interest image. The initial image and / or the target image may be processed in subsequent processes.

[0105] In some embodiments, the image acquired in 211 may be a 2D image or a 3D image. In some embodiments, the acquired image may be a grayscale image or a color image. In some embodiments, the acquired image may be a medical image, such as a CT image, an MRI image, a PET image, or any combination thereof.

[0106] In some embodiments, the image acquired at 211 may be provided to the decomposition module 202, the reconstruction module 204, the input / output device 150, the like, or any combination thereof. For example, the breast image acquired at 211 may be used to generate one or more sub-images by the decomposition module 202. In some embodiments, the image may be stored in the memory 140, an external data storage device via the network 160, any component capable of storage, the like, or any combination thereof.

[0107] At 212, the image acquired or generated at 211 may be decomposed. The decomposition may be performed by decomposition module 202. In some embodiments, the image may be decomposed into one or more layers. In some embodiments, decomposition module 202 may decompose the image using one or more decomposition algorithms. The decomposition algorithms may include bilateral filtering, wavelet filtering, Laplace transform, intrinsic image decomposition algorithms, or any combination thereof. In some embodiments, the image may be decomposed by different decomposition units. For example, the image may be decomposed by first decomposition unit 202-1 and / or second decomposition unit 202-2. In some embodiments, the image may be decomposed based on sub-images. For example, the first decomposition unit 202-1 may decompose the image into a first layer comprising low-frequency sub-images and high-frequency sub-images. The second decomposition unit 202-2 may decompose the low-frequency sub-images in the first layer into a second layer comprising low-frequency sub-images and high-frequency sub-images. Similarly, the low-frequency sub-images in the second layer may be further decomposed into an Nth layer by an Nth decomposition unit 202-N. In some embodiments, one or more sub-images obtained after decomposition may be downsampled. For example, if the decomposition module 202 decomposes the image using wavelet transform or Laplace transform, the obtained sub-images may be downsampled.

[0108] In some embodiments, the image may be decomposed into L layers by the decomposition module 202. For example, the image may be decomposed into a first layer by the first decomposition unit 202-1. The first layer may include low-frequency sub-images and high-frequency images. The low-frequency sub-images in the first layer may be further decomposed into a second layer by the first decomposition unit 202-1. The low-frequency sub-images in the second layer may be further decomposed into a third layer by the first decomposition unit 202-1. Similarly, the low-frequency sub-images in the (L-1)th layer may be further decomposed into the Lth layer by the first decomposition unit 202-1. Similarly, the image may be decomposed into L'+N layers by the decomposition module 202. For example, the image may be decomposed into L'+N layers by the second decomposition unit 202-2. In some embodiments, the image may be decomposed into L layers by two or more decomposition units (e.g., the first decomposition unit 202-1, the second decomposition unit 202-2, the Nth decomposition unit 202-N, etc.). For example, the image may be decomposed into a first layer by the first decomposition unit 202-1. The first layer may include low-frequency sub-images and high-frequency images. The low-frequency sub-image or high-frequency image in the first layer can be further decomposed into a second layer by the second decomposition unit 202-2. Similarly, the low-frequency sub-image or high-frequency image in the (L-1)th layer can be further decomposed into the Lth layer by the Nth decomposition unit 202-N. In some embodiments, two or more decomposition units can use the same decomposition algorithm or different decomposition algorithms. In some embodiments, two or more decomposition units can use the same or different parameters of the same decomposition algorithm.

[0109] In 213, the image acquired in 211 and / or the sub-image generated in 212 may be transformed. The transformation module 203 may perform 213. In some embodiments, one or more of the images (and / or sub-images) may be transformed by a grayscale value transformation unit 203-1 based on a grayscale value transformation curve. The grayscale value transformation curve may be determined by the grayscale value transformation unit 203-1 based on the grayscale values of elements in the image (and / or sub-image) to be transformed. In some embodiments, one or more images (and / or sub-images) may be transformed by a weight transformation unit 203-2 based on a weight map. The weight map may be determined by the weight transformation unit 203-2 based on the image (and / or sub-image) to be transformed. In some embodiments, in 212, one or more images (and / or sub-images) may be enhanced by an enhancement unit 203-3. For example, a histogram equalization technique may be used to enhance high-frequency sub-images. In some embodiments, one or more images (and / or sub-images) may be upsampled by an upsampling unit 203-4. For example, high-frequency sub-images may be upsampled by linear interpolation. In some embodiments, one or more images (or sub-images) can be replaced by another image (or sub-image). For example, a high-frequency sub-image (or low-frequency sub-image) in the xth layer (x is an integer) generated using a first decomposition algorithm can be replaced by a high-frequency sub-image (or low-frequency sub-image) in the xth layer generated using a second decomposition algorithm. In some embodiments, two or more transformed images (and / or sub-images) can be used to reconstruct a composite image in a subsequent process.

[0110] In 214, a composite image (or sub-image) can be reconstructed based on the two or more decomposed images (and / or sub-images) generated in 213. The reconstruction module 204 can perform 214. In some embodiments, two or more sub-images in the same layer can be used to reconstruct an image (or sub-image) in another layer. For example, a first layer including a low-frequency sub-image and a high-frequency sub-image can be used to reconstruct a composite image. For another example, a second layer including a low-frequency sub-image and a high-frequency sub-image can be used to reconstruct a low-frequency sub-image in the first layer. In some embodiments, the reconstructed low-frequency sub-image in the first layer together with the high-frequency sub-image in the first layer can be further used to reconstruct a composite image. In some embodiments, the reconstructed image can have enhanced information compared to the initial image. For example, image contrast and / or edge information are enhanced in the reconstructed image. For another example, noise information is reduced or removed in the reconstructed image. In some embodiments, the reconstructed image (or sub-image) in 214 may then be sent to and / or provided to the image acquisition module 201, the decomposition module 202, the transformation module 203, the memory 140, the input / output device 150, the network 160, or any combination thereof. For example, the reconstructed image may be displayed on the input / output device 150. For another example, the reconstructed image may be stored in the memory 140 or an external data storage device via the network 160.

[0111] Figure 3-A FIG. 2 is a schematic diagram of an exemplary image acquisition module 201 according to some embodiments of the present application. Figure 3-A As shown, the image acquisition module 201 may include an initial image acquisition unit 301 , a region of interest (ROI) extraction unit 302 , a region of interest edge extraction unit 303 , and a region of interest image determination unit 304 .

[0112] The initial image acquisition unit 301 can acquire an initial image. The initial image may include an image acquired via the network 160 from the imaging device 110, the memory 140, and / or an external data storage device. In some embodiments, the initial image may be a processed image generated by the decomposition module 202, the transformation module 203, and / or the reconstruction module 204. In some embodiments, a region of interest and / or a region of interest edge may be extracted from the initial image. In some embodiments, the initial image may be processed to generate a region of interest image. In some embodiments, the initial image may be decomposed by the decomposition module 202 and / or transformed by the transformation module 203. For example, a breast CT image may be acquired as the initial image, and the decomposition module 202 may decompose the initial image into one or more layers. In some embodiments, the initial image may be generated by an FFDM system or a DBT system. In some embodiments, the initial image may be designated as a positive image, wherein the grayscale values of background elements may be higher than the grayscale values of the region of interest. In some embodiments, the initial image may be designated as a negative image, wherein the grayscale values of background elements may be lower than the grayscale values of the region of interest.

[0113] The region of interest (ROI) extraction unit 302 can extract a region of interest from the initial image. In some embodiments, the region of interest may include a region in the image that corresponds to a tissue, an organ, a tumor, etc., or any combination thereof. For example, the region of interest may include a region in the image that corresponds to a breast region, a lung region, a bone region, a liver region, a brain region, a kidney region, any region of the body, etc., or any combination thereof.

[0114] For the sake of brevity, an image corresponding to an object (e.g., a tissue, organ, tumor, etc. of an object (e.g., a patient, etc.)) or a portion thereof (e.g., a region of interest in an image) may be referred to as the image of the object or a portion thereof, or an image including the object or a portion thereof (e.g., a region of interest), or the object itself. For example, a region of interest corresponding to an image of the liver may be described as the region of interest including the liver. For another example, an image including the liver may be referred to as an image of the liver, or simply as the liver. For the sake of brevity, a portion of an image corresponding to an object being processed (e.g., extracted, segmented, etc.) may be described as the processed object. For example, extracting a portion corresponding to the liver from the rest of the image may be described as extracting the liver.

[0115] In some embodiments, the region of interest may include a partial region of a tissue, organ, tumor, or any combination thereof. In some embodiments, the partial region may include a central region or a region near the center of a tissue, organ, tumor, or any combination thereof. In some embodiments, the partial region may include a region having elements with grayscale values within a predetermined range. In some embodiments, the region of interest may be extracted to provide a region of interest image. For example, a breast region may be extracted to generate a breast image.

[0116] The region of interest edge extraction unit 303 can extract the region of interest edge from the initial image. The region of interest edge can refer to the edge of a tissue, organ, or tumor of interest. For example, the region of interest edge can refer to the edge of a breast, lung, bone, liver, brain, kidney, or any other region of the body, or any combination thereof. In some embodiments, the extracted region of interest edge can be used to generate a region of interest image.

[0117] The region of interest image determination unit 304 may determine a region of interest image. The region of interest image may refer to an image having a region of interest and / or region of interest edges that can be extracted from the initial image. The region of interest image determination unit 304 may determine the region of interest image based on the region of interest extracted by the region of interest extraction unit 302 and / or the region of interest edges extracted by the region of interest edge extraction unit 303. In some embodiments, the region of interest image may include a breast image, a lung image, a bone image, a liver image, a brain image, a kidney image, any part of the body, or any combination thereof. For example, a breast image may be determined based on the breast region and the edges of the breast region. In some embodiments, the region of interest image may not include the background of the initial image. For example, the breast image may not include the background of the chest in the initial image. In some embodiments, the region of interest image may be decomposed by the decomposition module 202. For example, the breast image may be decomposed by the decomposition module 202 into a low-frequency sub-image and a high-frequency sub-image.

[0118] Figure 3-B FIG. 3 is a flow chart of an exemplary process of acquiring an image according to some embodiments of the present application. The process may include acquiring an initial image 311 , extracting a region of interest 312 , extracting a region of interest edge 313 , and determining a region of interest image 314 .

[0119] In 311, an initial image may be acquired by the initial image acquisition unit 301. In some embodiments, the initial image may be acquired from the imaging device 110, the memory 140, an external data storage device via the network 160, or any combination thereof. In some embodiments, the initial image may be a CT image, an MRI image, a PET image, an infrared image, or any combination thereof. For example, the imaging device 110 may generate a chest CT image, and the chest CT image may be acquired by the initial image acquisition unit 301 in 311. In some embodiments, the initial image may be transformed into a logarithmic domain image based on a logarithmic transformation. The logarithmic domain image may be processed in subsequent operations.

[0120] In 312, a region of interest may be extracted based on the initial image acquired in 311. Operation 312 may be performed by the region of interest extraction unit 302. In some embodiments, the region of interest may be extracted based on one or more segmentation algorithms, including, for example, OSTU technology, watershed algorithm, threshold segmentation, region growing segmentation, energy-based 3D reconstruction segmentation, level set-based segmentation, region segmentation and / or merging segmentation, edge tracking segmentation, statistical pattern recognition, C-means clustering segmentation, deformable model segmentation, graph search segmentation, neural network segmentation, geodesic minimum path segmentation, target tracking segmentation, atlas-based segmentation, rule-based segmentation, coupled surface segmentation, model-based segmentation, deformable biological segmentation, etc., or any combination thereof.

[0121] In some embodiments, the region of interest can be extracted based on the grayscale values of the initial image. In some embodiments, a grayscale histogram can be generated based on the initial image. In some embodiments, the region of interest can be extracted based on the grayscale histogram. In some embodiments, the segmentation algorithm can be determined based on the characteristics of the grayscale histogram. For example, the grayscale histogram of the initial breast image may have a bimodal pattern. Given that the OSTU technique can be relatively efficient and accurate for bimodal grayscale histograms, the OSTU technique can be used to segment the initial image. For another example, a watershed algorithm can be used to extract the breast region.

[0122] In some embodiments, the algorithm for extracting the region of interest may be stored in the region of interest extraction unit 302, the memory 140, the network 160, or other removable storage devices. Exemplary removable storage devices may include flash drives, floppy disks, optical disks, memory cards, compact disks, magnetic tapes, or any combination thereof. In some embodiments, the algorithm for extracting the region of interest may be retrieved from one or more external data storage devices via the network 160.

[0123] In 313, a region of interest edge (e.g., a breast edge) may be extracted based on the initial image acquired in 311. Operation 313 may be performed by the region of interest edge extraction unit 303. In some embodiments, the region of interest edge may be extracted based on grayscale value variation characteristics of the edge. In some embodiments, the grayscale values of elements at the region of interest edge may be similar to the grayscale values of the background. In some embodiments, the region of interest edge may exhibit variations in the grayscale values of elements at the region of interest edge. In some embodiments, before extracting the region of interest edge, the grayscale values of elements at the region of interest edge may be enhanced, thereby making the region of interest edge more distinguishable from the background and facilitating region of interest edge extraction. In some embodiments, if the initial image is an FFDM image, the initial image may have varying noise levels due to, for example, variations in X-ray dose. In some embodiments, if the grayscale values are enhanced, the noise level of the initial image may affect the grayscale values of elements at the region of interest edge. In some embodiments, to reduce or eliminate the effects of noise, denoising may be performed before enhancing the grayscale values of elements at the region of interest edge. In some embodiments, considering that the attenuation of X-rays may follow an exponential distribution, in order to reduce computational complexity and / or improve efficiency of ROI edge extraction, the ROI edge may be extracted based on a logarithmic domain image derived from the initial image.

[0124] In some embodiments, the log domain image can be denoised based on a denoising algorithm. The denoising algorithm may include a Gaussian filter algorithm, a mean filter algorithm, a non-local means (NLM) algorithm, a three-dimensional block technique (BM3D) algorithm, a total variation algorithm, a partial differential equation (PDE) algorithm, a wavelet threshold algorithm, or the like, or any combination thereof. In some embodiments, in 313, the denoised log domain image can be preprocessed by grayscale value transformation. In some embodiments, the grayscale value transformation may refer to a gradient transformation of the grayscale values of elements of the log domain image. In some embodiments, a Sobel gradient operator may be used in the gradient transformation. In some embodiments, a differential algorithm may be used in the grayscale value transformation. In other words, a differential operation can be performed on the denoised log domain image.

[0125] It should be noted that in some embodiments, if the initial image is acquired directly from the FFDM system, the grayscale values of the elements of the region of interest may be lower than the grayscale values of the background. That is, the grayscale values in different regions may have the following relationship: grayscale value of background elements > grayscale value of elements at the edge of the region of interest > grayscale value of elements at the region of interest. In some embodiments, after denoising of the logarithmic domain image, grayscale value fluctuations can be reduced (for example, the grayscale values of elements of the background may tend to be at the same level). In some embodiments, after grayscale value transformation, the grayscale values in different regions may have the following relationship: grayscale value of elements at the edge of the region of interest > grayscale value of elements at the region of interest > grayscale value of background elements. In some embodiments, by utilizing denoising and / or grayscale value transformation of the logarithmic domain image, it is possible to avoid extracting elements in the background as elements at the edge of the region of interest, and to improve the accuracy of the edge of the region of interest.

[0126] In some embodiments, after denoising and grayscale conversion, the region of interest edge can be extracted from the logarithmic domain image. In some embodiments, an iterative technique can be used in the region of interest edge extraction. In some embodiments, after denoising and grayscale conversion, the logarithmic domain image can be represented by floating-point data. When an iterative technique is used in the region of interest edge extraction, floating-point data can be used, thereby avoiding the need to convert the floating-point data to integer data and the loss of image precision or image quality associated with the conversion.

[0127] The iterative technique for determining the edge of the region of interest may include a first iteration, a second iteration, and / or an Nth iteration. The initial threshold T0 may be determined as half the sum of the maximum grayscale value and the minimum grayscale value in the denoised and / or preprocessed logarithmic domain image. In the first iteration, the denoised and / or preprocessed logarithmic domain image may be divided into a first region and a second region based on the threshold T0. The grayscale value of the elements in the first region may be greater than T0. The grayscale value of the elements in the second region may be equal to and / or less than T0. In some embodiments, the average grayscale value of the elements in the first region and the second region may be calculated respectively. For example, G first can represent the average gray value of the elements in the first region, and G second In some embodiments, T0 may be updated to have a value of T1, where T1 may be G first and G second Half of the sum (ie, T1 = (G first +G second ) / 2).

[0128] In the second iteration, if abs(T1-T0)>1 (abs may refer to an absolute value), the denoised and / or preprocessed logarithmic domain image may be further divided into a third region and a fourth region. The grayscale value of the elements in the third region may be greater than T1. The grayscale value of the elements in the fourth region may be equal to and / or less than T1. In some embodiments, the average grayscale value of the elements in the third region and the average grayscale value of the elements in the fourth region may be calculated respectively. For example, G third can represent the average gray value of the elements in the third region, and G fourth The average grayscale value of the elements of the fourth region may be represented. In some embodiments, T1 may be updated to have a value of T2, where T2 may be G third and G fourth Half of the sum (ie, T2 = (G third +G fourth ) / 2).

[0129] In the third iteration, if abs(T2-T1)>1, the denoised and / or preprocessed logarithmic domain image can be further divided into a fifth region and a sixth region. The grayscale value of the elements in the fifth region can be greater than T2. The grayscale value of the elements in the sixth region can be equal to and / or less than T2. In some embodiments, the average grayscale value of the elements in the fifth region and the average grayscale value of the elements in the sixth region can be calculated respectively. For example, G fifth can represent the average gray value of the elements in the fifth region, and G sixth The average grayscale value of the elements of the sixth region may be represented. In some embodiments, T2 may be updated to have a value of T3, where T3 may be G fifth and G sixth Half of the sum (ie, T3 = (G fifth +G sixth ) / 2). Until (T i+1 -T i )<1, the iteration can be stopped and the denoised and / or preprocessed logarithmic domain image can be not further divided. i+1 Determine the threshold for edge extraction of the region of interest. i+1 To divide the denoised and / or preprocessed logarithmic domain image. i+1 A set of elements is determined to belong to the edge of the region of interest. In some embodiments, the threshold values (e.g., T0, T1, T2, T3, ..., T iIn some embodiments, the weight value may be determined based on a threshold variation characteristic during iteration. In some embodiments, the threshold variation characteristic may refer to a variation between two or more thresholds, such as T0, T1, T2, T3, ..., T i+1 Etc. In some embodiments, the weight may be less than 1.

[0130] In some embodiments, in addition to the iterative technique, the OTSU algorithm or any other algorithm that can facilitate the extraction of the region of interest edge can be used to extract the region of interest edge. In some embodiments, the extracted region of interest edge can be provided to the region of interest image determination unit 304 in 314. In some embodiments, the extracted region of interest edge can be stored in the region of interest edge extraction unit 303, the memory 140, the network 160, or other mobile storage devices.

[0131] In 314, a region of interest image may be determined. Operation 314 may be performed by the region of interest image determination unit 304. In some embodiments, the region of interest image may be determined based on the region of interest extracted in 312 and / or the region of interest edge extracted in 313. In some embodiments, the region of interest image may be generated by combining the region of interest and the region of interest edge. In some embodiments, the generated region of interest image may include one or more isolated elements (e.g., one or more elements of the background are extracted as elements of the region of interest image). In some embodiments, the isolated elements may be removed from the region of interest image to enhance the quality of the region of interest image. In some embodiments, the region of interest image may be provided to the decomposition module 202, the reconstruction module 204, the memory 140, the input / output device 150, and / or the network 160. For example, the breast image may be decomposed into L layers by the decomposition module 202.

[0132] It should be noted that Figure 3-A and Figure 3-B The above description is merely illustrative and should not be construed as the only embodiment. Those skilled in the art, once they understand the basic principles of operation, may modify or alter the diagrams and / or flow charts without departing from the principles. Such modifications and alterations remain within the scope of the present application. For example, steps 312, 313, and / or 314 may be skipped, meaning that the initial image may be sent to the decomposition module 202, the transformation module 203, and / or the reconstruction module 204 for further processing.

[0133] Figure 4-A FIG. 4 is a flow chart of an exemplary process of decomposing an image according to some embodiments of the present application. The process may include a first decomposition operation 401 and a second decomposition operation 402 .

[0134] In 401, an image may be decomposed into L (L is an integer, and L ≥ 1) layers based on the first decomposition of the decomposition module 202. The image may include the initial image acquired in 311, the region of interest extracted in 312, the region of interest edge extracted in 313, the region of interest image determined in 314, a decomposed image, a transformed image, an image acquired by the imaging device 110, the memory 140, and / or the network 160, or any combination thereof. In some embodiments, the image may be decomposed by the first decomposition unit 202-1, the second decomposition unit 202-2, and / or the Nth decomposition unit 202-N.

[0135] In some embodiments, the image can be decomposed by wavelet transform, bilateral filtering, Fourier algorithm, discrete cosine transform, Laplace transform, any algorithm capable of decomposing an image, or the like, or any combination thereof. For example, a bilateral filter can be used to decompose an image into a low-frequency sub-image and a high-frequency sub-image. For another example, an image can be decomposed into L layers by Laplace transform. In some embodiments, a layer can include a low-frequency sub-image and a high-frequency sub-image. For example, an image can be decomposed into 3 layers by Laplace transform. First, the image can be decomposed into a first layer including a low-frequency sub-image and a high-frequency sub-image by Laplace transform; second, the low-frequency sub-image in the first layer can be further decomposed into a second layer including a low-frequency sub-image and a high-frequency sub-image by Laplace transform; third, the low-frequency sub-image in the second layer can be further decomposed into a third layer including a low-frequency sub-image and a high-frequency sub-image by Laplace transform.

[0136] In some embodiments, an image can be decomposed into L layers using a wavelet transform. For example, an image can be decomposed into three layers using a wavelet transform. First, the image can be decomposed into a first layer comprising low-frequency sub-images and high-frequency sub-images using a wavelet transform. Second, the low-frequency sub-images in the first layer can be further decomposed into a second layer comprising low-frequency sub-images and high-frequency sub-images using a wavelet transform. Third, the low-frequency sub-images in the second layer can be further decomposed into a third layer comprising low-frequency sub-images and high-frequency sub-images using a wavelet transform.

[0137] In 402, the image may be decomposed into L'+N layers (L' and N are integers, L ≥ 1 and N ≥ 1) based on the second decomposition of the decomposition module 202. In some embodiments, the image may be decomposed by the first decomposition unit 202-1, the second decomposition unit 202-2, and / or the Nth decomposition unit 202-N. In some embodiments, the image to be decomposed in 402 may be the same as the image to be decomposed in 401. In some embodiments, the image to be decomposed in 402 may be different from the image to be decomposed in 401. In some embodiments, the image to be decomposed in 402 may be a sub-image generated in 401. In some embodiments, the image may be decomposed by wavelet transform, bilateral filtering, Fourier transform, discrete cosine transform, Laplace transform, any algorithm capable of decomposing an image, or any combination thereof.

[0138] In some embodiments, an image can be decomposed into L'+N layers by wavelet transform. For example, if L'=3 and N=1, the image can be decomposed into 4 layers by wavelet transform. First, the image can be decomposed into a first layer including low-frequency sub-images and high-frequency sub-images by wavelet transform; second, the low-frequency sub-images in the first layer can be further decomposed into a second layer including low-frequency sub-images and high-frequency sub-images by wavelet transform; third, the low-frequency sub-images in the second layer can be further decomposed into a third layer including low-frequency sub-images and high-frequency sub-images by wavelet transform; fourth, the low-frequency sub-images in the third layer can be further decomposed into a fourth layer including low-frequency sub-images and high-frequency sub-images by wavelet transform.

[0139] The sub-images generated in 401 and / or 402 can be provided to the image acquisition module 201, the transformation module 203, the reconstruction module 204, the memory 140, the input / output device 150, the network 160, etc., or any combination thereof. For example, the sub-image can be transformed by the grayscale value transformation unit 203-1. For another example, the sub-image can be transformed by the transformation module 203-2. For another example, one or more elements of a sub-image can be replaced by elements of another sub-image.

[0140] It should be noted that different algorithms can be used to decompose different layers. In some embodiments, the number of layers generated in 402 (L'+N) can be greater than the number of layers generated in 401 (L). For example, an image can be decomposed into 4 layers in 402, while the same image can be decomposed into 3 layers in 401. In some embodiments, 401 or 402 can be skipped. In some embodiments, one or more operations can be added before or after 401 and / or 402. For example, a third decomposition operation can be added after 402.

[0141] Figure 4-B FIG. 1 is a schematic diagram of an exemplary L layer decomposed by a first decomposition unit according to some embodiments of the present application. Figure 4-B As shown, image 411 can be decomposed into L layers. Image 411 can be decomposed by decomposition module 202. For example, image 411 can be decomposed by first decomposition unit 202-1. In some embodiments, one or more decomposition algorithms can be used to decompose image 411. For example, a Laplace transform can be used to decompose the image. In some embodiments, each layer can include two or more sub-images. For example, a single layer can include a low-frequency sub-image, a mid-frequency sub-image, and / or a high-frequency sub-image.

[0142] In some embodiments, image 411 may be decomposed into a first layer including low-frequency sub-images 412 and high-frequency sub-images 413. Low-frequency sub-images 412 in the first layer may be decomposed into a second layer including low-frequency sub-images 414 and high-frequency sub-images 415. Low-frequency sub-images 414 in the second layer may be decomposed into a third layer including low-frequency sub-images 416 and high-frequency sub-images 417. Similarly, low-frequency sub-images (not shown) in the (L-1)th layer may be decomposed into an Lth layer including low-frequency sub-images 418 and high-frequency sub-images 419. In some embodiments, high-frequency sub-images in a layer may be decomposed into layers including low-frequency sub-images and high-frequency sub-images.

[0143] Figure 4-C FIG. 1 is a schematic diagram of an exemplary L'+N layer decomposed by a second decomposition unit according to some embodiments of the present application. Figure 4-C As shown, image 421 can be decomposed into L'+N layers. The image can be decomposed by decomposition module 202. For example, the image can be decomposed by second decomposition unit 202-2. In some embodiments, a decomposition algorithm can be used to decompose the image. For example, a wavelet transform can be used to decompose the image.

[0144] In some embodiments, image 421 may be decomposed into a first layer including a low-frequency sub-image 422 and a high-frequency sub-image 423. Low-frequency sub-image 422 in the first layer may be decomposed into a second layer including a low-frequency sub-image 424 and a high-frequency sub-image 425. Low-frequency sub-image 424 in the second layer may be decomposed into a third layer including a low-frequency sub-image 426 and a high-frequency sub-image 427. Similarly, a low-frequency sub-image (not shown) in the (L'-1)th layer may be decomposed into an Lth layer including a low-frequency sub-image 428 and a high-frequency sub-image 429. Low-frequency sub-image 428 in the Lth layer may be decomposed into an (L'+1)th layer including a low-frequency sub-image 430 and a high-frequency sub-image 431. Low-frequency sub-image 430 in the (L'+1)th layer may be decomposed into an (L'+2)th layer including a low-frequency sub-image (not shown) and a high-frequency sub-image (not shown). Similarly, a low-frequency sub-image (not shown) in the (L'+N-1)th layer may be decomposed into an (L'+N)th layer including a low-frequency sub-image 432 and a high-frequency sub-image 433. In some embodiments, a high-frequency sub-image in a layer may be decomposed into a layer including a low-frequency sub-image and a high-frequency sub-image.

[0145] In some embodiments, L' may be equal to L. For example, if L' is equal to 3, then L may be equal to 3. In some embodiments, N is an integer greater than or equal to 1. For example, N may be equal to 1. In some embodiments, each layer may include two or more sub-images. For example, a single layer may include a low-frequency sub-image, a mid-frequency sub-image, and / or a high-frequency sub-image.

[0146] Figure 5-A FIG. 2 is a schematic diagram of a grayscale value conversion unit 203-1 according to some embodiments of the present application. Figure 5-A As shown, the grayscale value transform unit 203 - 1 may include a characteristic curve determining unit 501 , a characteristic curve dividing unit 502 , a transformation curve segment determining unit 503 , a transformation curve generating unit 504 and a first updating unit 505 .

[0147] The characteristic curve determination unit 501 may determine a characteristic curve. In some embodiments, the characteristic curve may be determined based on a target image, which may include, for example, an initial image, a region of interest, an edge of a region of interest, an image of a region of interest, or a sub-image. For example, the characteristic curve may be determined based on a low-frequency sub-image. The characteristic curve may show the relationship between the distance and the corresponding grayscale value of an element (e.g., Figure 6(as shown). In some embodiments, the distance may refer to the distance between an element of the target image and a reference edge of the target image. In some embodiments, the reference edge of the target image may refer to the edge of a region of interest in the target image, or the outer edge of the entire target image. In some embodiments, the distance of an element of the target image may refer to the minimum distance from the element to the reference edge. For each element of the target image, the distance to the reference edge may be calculated. In some embodiments, a line with the minimum distance may be determined for each of multiple elements of the target image. In some embodiments, the value of the minimum distance may be represented by the number of elements in the line with the minimum distance. In some embodiments, an element may have one or more distances to the reference edge, and the distance in the characteristic curve may refer to the minimum distance. In some embodiments, different elements may have different distances (e.g., 0.5 cm, 0.8 cm, 1 cm, etc.); for example, elements farther from the reference edge may have a greater distance than elements closer to the reference edge. In some embodiments, the distance may correspond to a grayscale value. In some embodiments, one or more elements may have the same distance and different grayscale values; the grayscale value corresponding to the distance may then be determined as the average grayscale value of elements with the same distance. For example, three elements A, B, and C have the same minimum distance of 0.5 cm; then the grayscale value corresponding to the distance can be calculated as G 0.5cm =(G A +G B +G C ) / 3.

[0148] The characteristic curve dividing unit 502 may divide the characteristic curve into two or more segments. In some embodiments, the characteristic curve dividing unit 502 may generate N curve segments based on the characteristic curve, where N may be an integer greater than 1.

[0149] The transformation curve segment determination unit 503 may determine one or more transformation curve segments. A transformation curve segment may be a portion of a transformation curve. In some embodiments, a transformation curve segment may be determined based on one or more characteristic curve segments. For example, a transformation curve segment may be determined based on a corresponding characteristic curve segment.

[0150] The transformation curve generation unit 504 may generate a transformation curve based on one or more transformation curve segments. The transformation curve may show the relationship between the grayscale values of elements in the target image (or sub-image) before and after grayscale value transformation (e.g., grayscale value transformation performed by the grayscale value transformation unit 203-1).

[0151] The first updating unit 505 may update the grayscale values of one or more elements of the target image based on the transformation curve. In some embodiments, the first updating unit 505 may transform the grayscale values within a preset range. In some embodiments, the first updating unit 505 may compress or amplify the grayscale values within the preset range. For example, the first updating unit 505 may amplify a grayscale value of 200 to 300, thereby changing the grayscale value of an element with an original grayscale value of 200 to 300. For another example, the first updating unit 505 may reduce the grayscale value of 200 to 100, thereby changing the grayscale value of an element with an original grayscale value of 200 to 100.

[0152] Figure 5-B FIG. 1 is a flow chart of an exemplary process of transforming the grayscale value of an image according to some embodiments of the present application. Figure 5-B As shown, the process may include determining a characteristic curve 511 , dividing the characteristic curve 512 , determining transformation curve segments 513 , generating a transformation curve based on the determined transformation curve segments 514 , and updating the grayscale value of the element based on the transformation curve 515 .

[0153] In 511, a characteristic curve may be determined based on the target image. Operation 511 may be performed by characteristic curve determination unit 501. In some embodiments, a region of interest may be extracted as the target image. In some embodiments, a high-frequency sub-image or a low-frequency sub-image may be designated as the target image. A reference edge may be determined in the target image. In some embodiments, the edge of the region of interest may be determined as the reference edge. In some embodiments, the outer edge of the entire target image may be determined as the reference edge. In some embodiments, the reference edge may be determined automatically, semi-automatically, or manually. For example, a user may manually define the reference edge through the GUI of input / output device 150. For each element of the target image, a distance to the reference edge may be calculated. In some embodiments, a line with a minimum distance may be determined for each of multiple elements of the target image. In some embodiments, the value of the minimum distance may be represented by the number of elements in the line with the minimum distance. In some embodiments, one or more elements of the target image may have the same minimum distance value. The grayscale value corresponding to the minimum distance D may be the average grayscale value of the grayscale values of elements with the same or similar minimum distance D. For example, the grayscale value corresponding to the minimum distance D can be determined by dividing the sum of the grayscale values of elements with the same or similar value at the minimum distance D by the number of these elements. For example, a minimum distance of 1.5 cm (or multiple elements, such as 110, 115, 117, 120, etc.) can correspond to element a, element b, and element c. The grayscale value corresponding to 1.5 cm (or multiple elements, such as 110, 115, 117, 120, etc.) can be calculated by dividing the sum of the grayscale values of element a, element b, and element c by 3.

[0154] In 512, the characteristic curve determined in 511 can be divided into N characteristic curve segments. Operation 512 can be performed by the characteristic curve dividing unit 502. In some embodiments, the grayscale value of the elements of the reference edge can be similar to the grayscale value of the background. In some embodiments, one or more elements of the reference edge and its neighborhood can have a higher grayscale value than one or more elements of the internal area. In order to reduce the difference in grayscale value of the elements between the reference edge and the neighborhood and the internal area of the reference edge, the grayscale value of the elements of the reference edge and its neighborhood can be reduced. In some embodiments, the reference edge and its neighborhood can have a lower grayscale value than the internal area. In order to reduce the difference in grayscale value between the reference edge and the neighborhood and the internal area of the reference edge, the grayscale value of the elements of the reference edge and its neighborhood can be increased.

[0155] In some embodiments, the grayscale interval can be determined before the characteristic curve is divided into characteristic curve segments. As an example, the grayscale value of the elements of the reference edge and its neighborhood decreases, and the maximum value of the grayscale interval (such as Figure 7 Point A shown in ) can be determined as the maximum grayscale value of the elements of the reference edge before the grayscale values of the elements of the reference edge are modified. In some embodiments, the minimum value of the grayscale interval can be determined based on the various applications of the technology. In some embodiments, the maximum value of the grayscale interval and / or the minimum value of the grayscale interval can be adjusted based on the processing efficiency of the image processing system 130. In some embodiments, the minimum grayscale value can be a grayscale value corresponding to a preset distance in the characteristic curve. For example, the preset distance can be 2 cm. If it is necessary to transform the grayscale value of an element located within a predetermined distance from the reference edge, the minimum value of the grayscale interval (such as Figure 7 ) is determined as the corresponding grayscale value of the preset distance. In some embodiments, the minimum value of the grayscale interval can be set automatically, semi-automatically, or manually. For example, the user can define the minimum value through the GUI of the input / output device 150. For another example, the user can specify a distance from the reference edge through the GUI of the input / output device 150 to define the minimum value.

[0156] In some embodiments, after the grayscale interval is determined, the portion of the characteristic curve within the grayscale interval can be divided into N segments. The number of characteristic curve segments (N) can be determined based on the number of grayscale values greater than the minimum value of the determined grayscale interval. In some embodiments, since the grayscale value corresponding to the distance can be determined as the average grayscale value of elements with the same distance, the grayscale values in the characteristic curve can be discrete data points. That is, the number N can be a determined number. For example, if the minimum value of the grayscale interval is determined to be 200, the grayscale values greater than 200 can include 300, 400, 500, 600, and 700. Therefore, the number of grayscale values greater than 200 is 5, and based on the number of grayscale values, the characteristic curve can be divided into 5 segments.

[0157] In 513, a transformation curve segment can be determined based on the N divided characteristic curve segments generated in 512. Operation 513 can be performed by the transformation curve segment determination unit 503. In some embodiments, the transformation curve segment can correspond to the characteristic curve segment. In some embodiments, the transformation curve segments can be determined one by one. In some embodiments, the second transformation curve segment can be determined based on the previously determined first transformation curve segment. In some embodiments, the order of different transformation curve segments can be determined based on the grayscale intervals of different characteristic curve segments. For example, a characteristic curve segment with a relatively high grayscale interval can be used first to determine its corresponding transformation curve segment, and then a characteristic curve segment with a relatively low grayscale interval can be used to determine its corresponding transformation curve segment. For another example, a characteristic curve segment with a relatively low grayscale interval can be used first to determine its corresponding transformation curve segment, and then a characteristic curve segment with a relatively high grayscale interval can be used to determine its corresponding transformation curve segment. In some embodiments, the slope of the transformation curve segment can be determined based on the corresponding characteristic curve segment.

[0158] Figure 7 Shown is an exemplary characteristic curve segmented into a plurality of characteristic curve segments. Figure 7 The gray value of the characteristic curve segment (from point A to the preset point P) can be found in [G A , G P ], where G A It can represent the gray value corresponding to the distance of point A in the figure, G P It can represent the gray value corresponding to the distance of point P in the figure. A It can represent the maximum gray value of the characteristic curve segment, and G P The minimum grayscale value of a characteristic curve segment can be represented. As shown in the figure, the characteristic curve can be segmented into three characteristic curve segments, including segments AB, BC, and CP. The transformation curve segment corresponding to segment AB can be determined first, and then segments BC and CP can be determined, or vice versa. It should be noted that the number of segments is for illustrative purposes only and is not intended to limit the scope of this application.

[0159] With reference to characteristic curve segment AB, the slope of transformation curve segment A'B' corresponding to characteristic curve segment AB can be determined based on the grayscale value of point P in the characteristic curve and the grayscale values of points A and B in characteristic curve segment AB. For example, the slope of transformation curve segment A'B' can be determined as the ratio of the grayscale value of point P in the characteristic curve to the average of the grayscale values of points A and B in characteristic curve segment AB, as shown in Formula 1:

[0160]

[0161] where K A’B’ represents the slope of the transformation curve segment A'B', G P Represents the gray value of point P, G A Represents the grayscale value of point A, G B represents the grayscale value of point B. The slopes of other transformation curve segments (e.g., segment B'C' corresponding to characteristic curve segment BC, segment C'P' corresponding to characteristic curve segment CP, etc.) can be calculated in a similar manner to segment A'B'. In some embodiments, the initial point of transformation curve segment A'B' can be determined as point A'. In some embodiments, point A' can have the same grayscale value as point A in the characteristic curve. In some embodiments, the slope K can be calculated based on the slope K. A’B’ And the gray value of point A' is used to determine the transformation curve segment A'B'.

[0162] Regarding the transformed curve segment B'C' corresponding to the characteristic curve segment BC, the slope of segment B'C' can be determined similarly to how the slope of segment A'B' corresponding to the characteristic curve segment AB is determined. In some embodiments, the initial point of the transformed curve segment B'C' can be determined as point B'. The grayscale value of point B' can be determined according to equation (2):

[0163]

[0164] Among them G B’ Indicates the gray value of point B', G A’ Indicates the gray value of point A', d A Indicates the distance from the initial point (point A) in the characteristic curve segment AB, d B Indicates the distance from the end point (point B) in the characteristic curve segment AB, d P Indicates the distance from the preset point P, G A Represents the grayscale value of the initial point (point A) of the characteristic curve segment AB, G P In some embodiments, the gray value of the preset point P can be based on the slope K B’C’ and the gray value of point B' (ie, G B’) to determine the transformation curve segment B'C'. Therefore, each transformation curve segment corresponding to the characteristic curve segment can be determined as described above.

[0165] In some embodiments, the slope of the Nth transformation curve segment may be the ratio of the grayscale value of a preset point P in the characteristic curve to the average grayscale value of the initial point and the end point of the Nth characteristic curve segment. The grayscale value of the initial point of the Nth transformation curve segment may be the sum of the grayscale value of the initial point in the (N-1)th transformation curve segment and the grayscale change of the corresponding (N-1)th characteristic curve segment. Accordingly, the grayscale change of the (N-1)th characteristic curve may refer to the grayscale change of the initial point and the end point of the (N-1)th characteristic curve. The grayscale value of the initial point of the Nth transformation curve segment may be determined according to equation (2), where G B’ It can correspond to the gray value of the initial point of the Nth transformation curve segment, G A’ It can correspond to the gray value of the initial point of the (N-1)th transformation curve segment, d A The distance d may correspond to the initial point in the (N-1)th characteristic curve segment. B The distance d may correspond to the end point in the (N-1)th characteristic curve segment. P The distance corresponding to the preset point P, G A It can correspond to the gray value of the initial point of the (N-1)th characteristic curve segment, G P It can correspond to the grayscale value of the preset point P.

[0166] In some embodiments, in 513, a point (eg, a preset point P) having the minimum value of the grayscale interval (determined in 512) may be designated as the initial point. Figure 7 As shown, point P can be designated as the initial point of segment CP, point C can be designated as the initial point of segment BC, and point B can be designated as the initial point of segment AB. Therefore, the initial point of transformation curve segment C'P' can be designated as point P', the initial point of transformation curve segment B'C' can be designated as point C', and the initial point of transformation curve segment A'B' can be designated as point B'. In some embodiments, the grayscale value of point P' can be the same as the grayscale value of point P, and the grayscale values of points C' and B' can be determined based on equation (2). For example, the grayscale value of point C' can be determined according to equation (3):

[0167]

[0168] Among them G C’ Indicates the gray value of the initial point C', G P’ Indicates the gray value of point P', d P Indicates the distance from the initial point P in the characteristic curve segment CP, d C Indicates the distance from the end point C in the characteristic curve segment CP, dA represents the distance to point A, G P represents the gray value of point P in the characteristic curve segment CP, and G A Represents the grayscale value of point A in the midpoint of the characteristic curve segment.

[0169] In some embodiments, the slopes of transformation curve segment C'P' and segment B'C' can be determined according to equation (1). Transformation curve segment C'P' corresponding to segment CP can be determined based on the grayscale value of point P' and the slope of C'P'. Thus, transformation curve segment B'C' and segment A'B' can be determined in this manner.

[0170] In 514, a transformation curve may be generated based on the transformation curve segments determined in 513. Operation 514 may be performed by the transformation curve generation unit 504. In some embodiments, the transformation curve may be generated by curve fitting based on the determined transformation curve segments. In some embodiments, the transformation curve may be generated using one or more fitting techniques, including, for example, least squares techniques, Lagrange interpolation techniques, Newton iteration techniques, cubic spline interpolation techniques, or any combination thereof. For example, the transformation curve may be generated by curve fitting based on Lagrange interpolation techniques. In some embodiments, the transformation curve may be used to update the grayscale values of elements in the target image.

[0171] In 515, the grayscale values of one or more elements in the target image may be updated. Operation 515 may be performed by the first updating unit 505. In some embodiments, the grayscale values may be updated based on the transformation curve generated in 514. In some embodiments, the grayscale values of elements whose grayscale values are within the grayscale interval determined in 512 may be updated. In some embodiments, the grayscale values may be compressed or amplified based on the transformation curve. In some embodiments, the transformation curve may be used to compress or amplify the grayscale values of a reference edge in the low-frequency sub-image and / or the high-frequency sub-image. In some embodiments, after the update, the grayscale values of elements within the grayscale interval may be modified (decreased or increased). In some embodiments, the modification effect may be linear or nonlinear. In some embodiments, the transformation curve may be used to modify the grayscale values of elements whose distance from the reference edge is within a preset range. In some embodiments, by adjusting the appropriate grayscale interval determined in 512, the grayscale values of background elements in the target image may not be updated. In some embodiments, after the update, the edge of the region of interest may be distinguished from the background.

[0172] In some embodiments, the grayscale values of one or more elements of the target image can be compressed based on a transformation curve. The transformation curve can be determined based on a positive image, in which the grayscale values of background elements in the positive image are higher than the grayscale values of the region of interest. In some embodiments, the minimum value of the grayscale interval of the transformation curve can be the grayscale value of a predetermined point. The maximum value of the grayscale interval of the transformation curve can be the maximum grayscale value of an element of a reference edge. In some embodiments, the number of segments (N) can be determined based on the number of grayscale values greater than or equal to the minimum value of the grayscale interval.

[0173] In some embodiments, the grayscale values of one or more elements of the target image can be amplified based on a transformation curve. The transformation curve can be determined based on a negative image, in which the grayscale values of background elements in the negative image are lower than the grayscale values of the region of interest. The grayscale values of one or more elements of the negative image can be higher after being amplified based on the transformation curve than before the transformation. For example, the grayscale values of one or more elements of a reference edge can be amplified. The grayscale values of background elements may not be adjusted. Therefore, the contrast between the reference edge and the background can be enhanced. The enhanced contrast makes it easier for a user (e.g., a doctor) to distinguish between the region of interest and the background.

[0174] As an example, a transformation curve is used to amplify the grayscale value of one or more elements of a target image, and the grayscale value of the initial point and slope of the transformation curve segment can be determined according to equation (1), equation (2) and / or equation (3). In some embodiments, the minimum value of the grayscale interval of the transformation curve can be the minimum grayscale value of the element of the reference edge. The maximum value of the grayscale interval of the transformation curve can be the grayscale value of a preset point. In some embodiments, the number of transformation curve segments (N) can be determined based on the number of grayscale values that are less than or equal to the maximum value of the grayscale interval.

[0175] Figure 6 is a schematic diagram of an exemplary characteristic curve according to some embodiments of the present application. As described above, the characteristic curve can represent the relationship between the distance and the corresponding grayscale value. In some embodiments, the distance can be represented by the number of elements between the element in the target image and the reference edge of the target image. The horizontal axis of the characteristic curve can represent the minimum distance between multiple elements of the target image and the reference edge. The minimum distance is displayed as the number of elements. The vertical axis of the characteristic curve can represent the corresponding average grayscale value of one or more elements that may have the same distance. It should be understood that Figure 6 The specific characteristic curves shown in are for illustrative purposes and are not intended to limit the scope of the present application. Figure 6 The characteristic curve shown in can be determined from the low-frequency sub-image. According to the characteristic curve, the grayscale values corresponding to different distances can be determined.

[0176] Figure 7FIG is a schematic diagram of an exemplary characteristic curve divided into multiple characteristic curve segments according to some embodiments of the present application. Figure 7 As shown, the characteristic curve can be divided into characteristic curve segments. In some embodiments, the grayscale values shown in the vertical axis can be discrete, and the continuous curve is only an approximation for illustration purposes. Figure 7 In the example, point P can be predetermined. The grayscale interval (where the grayscale value can be transformed) can be determined as [G P , G A ], where G P It can represent the lower limit of the grayscale range, and G A It can represent the upper limit of the grayscale range. [G P , G A The characteristic curve within the range can be divided into three segments: segment AB, segment BC and segment CP. It should be noted that Figure 7 The number of characteristic curve segments shown in is for illustrative purposes only and is not intended to limit the scope of the present application.

[0177] Figure 8 is a schematic diagram of exemplary transformation curves according to some embodiments of the present application. Figure 8 shows the grayscale intervals in which the grayscale values can be transformed. P Can be the minimum value of the grayscale interval, and G P Can correspond to Figure 7 The grayscale value of the preset point P is shown. A can be the maximum value of the grayscale interval, and G A Can correspond to Figure 7 Grayscale value of midpoint A. Figure 8 The transformation curve P'A' shown can be based on Figure 7 The transformation curve P'A' can show the effect of the correction on the grayscale values. Figure 8 The conversion curve shown in the example shows that [G P , G A ] is the grayscale value within .

[0178] Figure 9-A FIG. 2 is a schematic diagram of an exemplary weight transformation unit 203 - 2 according to some embodiments of the present application. The weight transformation unit 203 - 2 may include a weight map determining unit 901 and a second updating unit 902 .

[0179] The weight map determination unit 901 can determine a weight map for a target image (or sub-image). The target image (or sub-image) can refer to an image (or sub-image) to be processed or being processed by the image processing system 130. The target image (or sub-image) can include a high-frequency sub-image, a low-frequency sub-image, a grayscale image, a color image, etc. In some embodiments, the weight map determination unit 901 can generate a weight map based on the target image (or sub-image). In some embodiments, the weight map determination unit 901 can obtain the weight map from the memory 140 or an external data storage device via the network 160. The weight map can refer to a 2D image or a 3D image, in which the value of each element represents the weight of the corresponding element of the target image (or sub-image). In some embodiments, the weight map can have the same size as the target image (or sub-image). In some embodiments, the value of an element at a position in the weight map can be the weight of the element at the corresponding position in the target image (or sub-image). For example, the weight of an element at position (x, y) or a voxel at position (x, y, z) of a target image (or sub-image) may be the value of the element at position (x, y) or the voxel at position (x, y, z) of the weight map.

[0180] The second updating unit 902 can update the values of one or more elements of the target image (or sub-image) through the image processing system 130. The value of an element can refer to information of the element, including, for example, grayscale value, brightness, color, etc., or any combination thereof. In some embodiments, the second updating unit 902 can update the values of one or more elements based on the weight map determined by the weight map determination unit 901. In some embodiments, the second updating unit 902 and the first updating unit 505 can be integrated into a single updating unit having the functions of both units.

[0181] Figure 9-B 1 is a flowchart of an exemplary process for transforming a target image based on a weight map according to some embodiments of the present application. As shown in the present application, the exemplary process for transforming a target image based on a weight map may include determining a weight map and updating the target image based on the weight map.

[0182] At 911, a weight map may be determined based on the target image (or sub-image). Weight map determination 911 may be performed by weight map determination unit 901. In some embodiments, the weight map may be determined based on the grayscale values of elements of the target image (or sub-image). In some embodiments, the target image (or sub-image) may include first-category elements and / or second-category elements. In some embodiments, first-category elements may refer to noise elements in the target image, while second-category elements may refer to other elements in the target image other than noise elements. Noise elements may refer to pixels / voxels with noise, including, for example, Gaussian noise, salt and pepper noise, shot noise, quantization noise, film grain noise, anisotropic noise, etc., or any combination thereof. Other elements may refer to pixels / voxels in the target image other than noise pixels / voxels. In some embodiments, elements with different grayscale values may have different weights. In some embodiments, the weights of elements with different grayscale values may be determined using different techniques. For example, first-category elements may have lower grayscale values than second-category elements, and their weights may be determined using different techniques than those for second-category elements. In some embodiments, the weight map determination unit 901 may determine the grayscale interval of the first type of element in the target image (or sub-image). In some embodiments, the weight map determination unit 901 may adjust the determined grayscale interval of the first type of element in the target image (or sub-image). In some embodiments, the weight map determination unit 901 may determine the grayscale interval of the second type of element in the target image (or sub-image). In some embodiments, if the determined grayscale interval of the first type of element is adjusted, the grayscale interval of the second type of element may be adjusted accordingly. For example, for a grayscale interval of [G1, G N ] target image, where the grayscale interval of the first type of elements is determined as [G1, G x ], and the grayscale interval of the first type of elements is determined as [G x , G N ], if the grayscale interval of the first type of elements is adjusted to [G1, G x / 2], then the grayscale interval of the first type of elements can be adjusted to [G x / 2, G N In some embodiments, the grayscale interval of the first type of element can be mapped to a first range, and the grayscale interval of the second type of element can be mapped to a second range. Figure 9-C An exemplary process for determining a weight map is shown.

[0183] In 912, the target image (or sub-image) may be updated based on the weight map determined in 911. Updating the target image (or sub-image) in 912 may be performed by the second updating unit 902. In some embodiments, the target image (or sub-image) may be updated based on the product of the grayscale value of an element of the target image (or sub-image) and the corresponding weight of the weight map. In some embodiments, the update in 912 may be performed element by element. For example, the value of an element (x, y) (or (x, y, z)) of the target image (or sub-image) may be updated by multiplying the value itself by the weight of the same element (x, y) (or (x, y, z)) of its weight map.

[0184] In some embodiments, as Figure 4-B and / or Figure 4-C The high-frequency sub-image of the image layer shown in can be transformed based on a weight transformation. First, a weight map can be determined based on the high-frequency sub-image in 911. Second, the high-frequency sub-image can be updated based on the determined weight map obtained in 912. In some embodiments, an image reconstruction operation can be added after 912. For example, the updated high-frequency sub-image and the corresponding low-frequency sub-image can be used for reconstruction, and a transformed image (or image layer) can be obtained.

[0185] Figure 9-C is a flowchart of an exemplary process for determining a weight map according to some embodiments of the present application. At 921, the grayscale interval of the first-category element in the target image (or sub-image) can be determined. Operation 921 can be performed by the weight map determination unit 901. In some embodiments, the target image (or sub-image) can refer to a high-frequency sub-image, a low-frequency sub-image, a grayscale image, a color image, etc. In some embodiments, the grayscale interval of the first-category element can be determined based on a grayscale threshold. The grayscale threshold can be used to distinguish between first-category elements and second-category elements. In some embodiments, the grayscale threshold can be determined based on the default setting of the image processing system 130. For example, the grayscale threshold can be pre-stored in the memory 140, and the weight map determination unit 901 can retrieve the grayscale threshold from the memory 140 at 921. In some embodiments, the grayscale threshold can be determined manually or semi-manually. For example, a user can input or select a threshold from a list suggested by the image processing system 130 through a graphical user interface in the input / output device 150 to determine the grayscale threshold.

[0186] In some embodiments, the grayscale threshold can be determined based on one or more grayscale values of the target image (or sub-image). In some embodiments, the grayscale threshold can be determined based on the average grayscale value of all or part of the elements of the target image. average , to determine the grayscale threshold. For example, the average grayscale value, G, can be obtained by dividing the sum of the grayscale values of all elements in the target image by the number of all elements in the target image.average In some embodiments, the grayscale threshold can be determined based on the modified average grayscale value. For example, the grayscale threshold can be equal to the average grayscale value, G average Multiply by a preset coefficient k. In some embodiments, the coefficient k can be determined based on the grayscale value difference between an edge (e.g., an edge of a region of interest, an edge with a higher contrast than the background, an edge of the entire target image, etc.) and the first type of element of the target image. By way of example only, the preset coefficient k can be set within the range of [1, 3] (e.g., k = 3). It should be noted that the exemplary range of the coefficient k is provided for illustrative purposes only and is not intended to limit the scope of this application.

[0187] In some embodiments, the grayscale interval of the first type of elements can be determined according to a grayscale threshold. In some embodiments, in the high frequency sub-image, the first type of elements may have a grayscale value lower than the second type of elements, and the grayscale value of the element is lower than the grayscale threshold (e.g., G average itself or G average Multiplied by a preset coefficient k) can be regarded as a first-class element (eg, a noise element). For example, the grayscale value falls within [0, k×G average ] can be considered to belong to the first category.

[0188] In 922, the grayscale intervals of the first category elements in the target image (or sub-image) may be modified. Operation 922 may be performed by the weight map determination unit 901. In some embodiments, for example, in a high frequency sub-image, one or more regions of the target image (or sub-image) may have elements with weak details. In some embodiments, the elements with weak details may refer to elements or regions of the target image with a relatively high noise level. In some embodiments, the grayscale intervals of the elements with weak details may partially overlap with the grayscale intervals of the first category elements. In some embodiments, the overlapping grayscale intervals may be narrow, and it may be difficult to determine the modified grayscale threshold for distinguishing between the first category elements and the elements with weak details. In some embodiments, the grayscale intervals may be modified (e.g., [0, k×G average ]) to modify the grayscale interval of the first type of elements (for example, [0, k×G average ]). In some embodiments, the modification may refer to stretching or compressing the grayscale interval. The modification may include linear modification and / or nonlinear modification. In some embodiments, the weight map determination unit 901 may use one or more functions (e.g., sine function, logarithmic function, etc.) to nonlinearly modify the grayscale interval (e.g., [0, k×G average ]). In some embodiments, after modification, a modified grayscale interval (eg, a modified grayscale interval of [0, k×G average ]).

[0189] In 923, the grayscale intervals of the first type elements and / or the grayscale intervals of the second type elements may be adjusted based on the modified grayscale intervals obtained in 922. Operation 923 may be performed by the weight map determining unit 901. In some embodiments, the grayscale intervals of the first type elements may be adjusted based on the modified grayscale intervals (e.g., the modified [0, k×G average ]) to determine the first threshold. For example, the first threshold can be determined as the average grayscale value of the first type of elements in the modified grayscale interval (for example, the modified [0, k×G average ]). The first threshold value can be regarded as an updated grayscale threshold value for distinguishing between the first category elements and the second category elements. The grayscale interval of the first category elements and the grayscale interval of the second category elements can then be adjusted. For example, the grayscale interval of the first category elements can be [0, first threshold value], and the grayscale interval of the second category elements can be (first threshold value, second threshold value). The second threshold value can refer to the maximum grayscale value within the target image (or sub-image). In some embodiments, elements with a grayscale value equal to the first threshold value can be considered to belong to the first category. In some embodiments, elements with a grayscale value equal to the first threshold value can be considered to belong to the second category. In the following description, it is assumed that elements with a grayscale value equal to the first threshold value can be considered to belong to the first category. This assumption and the corresponding description are for illustrative purposes and are not intended to limit the scope of this application.

[0190] In 924, the weights of the first-category elements may be determined based on the grayscale interval of the first-category elements adjusted in 923 (e.g., [0, first threshold]). Operation 924 may be performed by the weight map determination unit 901. In some embodiments, the adjusted grayscale interval [0, first threshold] may be mapped to a range of [0, 1]. The mapping process may be performed based on one or more linear or nonlinear algorithms. The linear mapping process may be based on one or more linear algorithms. For example, all grayscale values within the grayscale interval [0, first threshold] may be divided by the first threshold, so that the grayscale interval [0, first threshold] may be mapped to [0, 1]. The nonlinear mapping process may be based on one or more nonlinear algorithms, including, for example, a sine function, a logarithmic function, etc., or any combination thereof. In some embodiments, the mapping range [0, 1] may be defined as the weight of the first-category element. That is, the grayscale value Gx(G x ∈[0, first threshold]) can be transformed into G after mapping x '(G x '∈[0,1]), and the value G x ' can be defined as having a gray value G x The weight of the first type of elements.

[0191] In 925, the weights of the second type elements may be determined based on the grayscale intervals of the second type elements adjusted in 923 (e.g., [first threshold, second threshold]). Operation 925 may be performed by the weight map determination unit 901. In some embodiments, the adjusted grayscale interval [first threshold, second threshold] may be mapped to a range of (1, G]. The mapping process may be performed based on one or more linear or nonlinear algorithms. The linear mapping process may be based on one or more linear algorithms. The nonlinear mapping process may be based on one or more nonlinear algorithms, including, for example, a sine function, a logarithmic function, etc., or any combination thereof. In some embodiments, the mapping range (1, G) may be defined as the weights of the second type elements. That is, the grayscale value G y (G y ∈(first threshold, second threshold]) can be transformed into G after mapping y '(G y '∈(1,G]), and the value G y ' can be defined as having a gray value G y The weight of the second type of elements in FIG. In some embodiments, the value of G can be determined or selected based on the desired image enhancement effect. For example, G can be determined to be 2, 3, etc. It should be noted that the value of G is for illustrative purposes only and is not intended to limit the scope of this application.

[0192] At 926, the grayscale values of the elements within the adjusted grayscale interval of the first-category elements obtained at 923 may be modified. Operation 926 may be performed by the weight map determining unit 901. In some embodiments, the grayscale values of the first-category elements may be modified based on the corresponding weights defined at 924. For example only, the grayscale values of the first-category elements may be replaced by the corresponding weights.

[0193] In 927, the grayscale values of the elements within the adjusted grayscale interval of the second-category elements obtained in 923 may be modified. Operation 927 may be performed by the weight map determining unit 901. In some embodiments, the grayscale values of the second-category elements may be modified based on the corresponding weights defined in 925. For example only, the grayscale values of the second-category elements may be replaced by the corresponding weights.

[0194] In 928, a weight map may be generated. Operation 928 may be performed by the weight map determination unit 901. In some embodiments, the weight map may be generated based on the grayscale values of the first-category elements modified in 926 and the grayscale values of the second-category elements modified in 927. By way of example only, after the modification processes in 926 and 927, the target image (or sub-image) may spontaneously become a weight map. In some embodiments, 926 and 927 may be omitted, and the weight map may be generated based on the weights of the first-category elements and the second-category elements and the corresponding positions of these weights. For example, the first-category elements may have a position (m, n) in the target image (or sub-image) and have an adjusted grayscale value G x (G x ∈[0, first threshold]), then the weight G of the element at the same position (m, n) in the weight graph can be given x '(G x '∈[0,1]). For another example, another element can have position (m',n') in the target image (or sub-image) and have the adjusted grayscale value G y (G y ∈(first threshold, second threshold]), then the weight G of the element at the same position (m', n') in the weight graph can be given y '(G y '∈(1,G]).

[0195] It should be noted that the above description and related flow charts about weight transformation unit 203-2 are merely examples and should not be understood as the only embodiment. For those skilled in the art, after understanding the basic principles of the connection between different units / operations, the connection between the units / operations and the units / operations can be modified or changed without departing from the principles. These modifications and changes are still within the scope of the above-mentioned current application. For example, 924 and 925 can be performed simultaneously or integrated into one operation. For another example, 926 and 927 can be omitted. For another example, 922 and 923 can be integrated into one operation. For another example, 924 and 926 can be integrated into one operation. For another example, 925 and 927 can be integrated into one operation.

[0196] Figure 10is a schematic diagram illustrating an exemplary process for generating a weight map. A target image 1001 (e.g., a high-frequency sub-image) has a plurality of first-category elements 1002, as indicated by the hollow circles in region 1010. Region 1010 may refer to a region of first-category elements 1002 within the grayscale interval determined in 921. Regions other than region 1010 may refer to regions of second-category elements. See region 1020 within 1001 but outside of region 1010. Assuming the grayscale interval of the first-category elements 1002 in region 1010 is [0, 100], after the nonlinear modification process described in 922, the grayscale interval of the first-category elements 1002 in region 1010 can be modified to [0, 200]. Assuming the average grayscale value of the first-category elements 1002 in region 1010 is 150, which is designated as a first threshold, after the adjustment process in 923, the grayscale interval of the first-category elements 1002 can be adjusted to [0, 150]. And the grayscale interval of the second type of elements can be adjusted to (150, second threshold]. The first type elements 1005 in the region 1040 with grayscale values in the range of [0,150] can be regarded as true first type elements. The grayscale interval [0,150] is mapped to [0,1] in 924. For example, all grayscale values in the range [0,150] can be divided by 150. The values in the mapping range [0,1] can represent the weights of the first type elements 1005 in the region 1040. The grayscale interval (150, second threshold] is mapped to (1,G) in 925. The values in the mapping range (1,G) can represent the weights of the second type of elements of the target image 1001 except the region 1040. In some embodiments, the value of G can be determined based on the second threshold. In some embodiments, the second threshold can refer to The maximum grayscale value within the target image 1001. In some embodiments, the value of G can be determined or selected based on the desired image enhancement effect. For example, G can be determined to be 2, 3, etc. It should be noted that the value of G is for illustrative purposes only and is not intended to limit the scope of this application. The values within the mapping range (1, G] can represent the weights of the second-category elements other than the first-category elements 1005 in the region 1040. A weight map corresponding to the target image 1001 can be obtained by replacing the grayscale values of the first-category elements 1005 in the region 1040 with corresponding weights in the range [0, 1], and replacing the grayscale values of the second-category elements other than the first-category elements 1005 in the region 1040 having corresponding weights in the range (1, G]. The weight map can show the weight of each element in the target image 1001.

[0197] Figure 11-A FIG. 1 is a flowchart of an exemplary process of reconstructing a composite image based on layers of a target image according to some embodiments of the present application. The reconstruction process may include acquiring a sub-image 1101 and reconstructing a composite image 1102 .

[0198] At 1101, a low-frequency sub-image and a high-frequency sub-image of a target image may be acquired. Operation 1101 may be performed by the reconstruction module 204. In some embodiments, the low-frequency sub-image and / or the high-frequency sub-image may be generated by the decomposition module 202. In some embodiments, the low-frequency sub-image and the high-frequency sub-image may be generated using the same decomposition algorithm. In some embodiments, the low-frequency sub-image and the high-frequency sub-image may be generated using different decomposition algorithms. For example, the low-frequency sub-image may be generated using a first decomposition algorithm, and the high-frequency sub-image may be generated using a second decomposition algorithm. For example, the low-frequency sub-image may be generated using a Laplace transform, while the high-frequency sub-image may be generated using a wavelet transform.

[0199] In some embodiments, the low-frequency sub-image and the high-frequency sub-image can be generated from the same target image (or sub-image). In some embodiments, the low-frequency sub-image and the high-frequency sub-image can be generated from the same layer. In some embodiments, the low-frequency sub-image and the high-frequency sub-image can be generated from different target images (or sub-images) or different layers of the same target image.

[0200] In some embodiments, the low-frequency sub-image may have been transformed by the transformation module 203. In some embodiments, the high-frequency sub-image may have been transformed by the transformation module 203. For example, the low-frequency sub-image and / or the high-frequency sub-image may have been transformed based on the following example: Figure 5-B For another example, the low-frequency sub-image and / or the high-frequency sub-image may have been transformed based on Figure 9-C For another example, the low-frequency sub-image and / or the high-frequency sub-image can be transformed as shown in FIG. Figure 2-A and Figure 2-B For example, the low-frequency sub-image and / or the high-frequency sub-image may be enhanced linearly or nonlinearly as shown. For another example, denoising may be performed on the low-frequency sub-image and / or the high-frequency sub-image. For another example, the low-frequency sub-image and / or the high-frequency sub-image may have been transformed by one or more interpolation processes (e.g., the sub-image may have been upsampled by upsampling unit 203-4). In some embodiments, the low-frequency sub-image and / or the high-frequency sub-image may be transformed by one or more transformation processes described herein.

[0201] In some embodiments, the (transformed) low-frequency sub-image and the (transformed) high-frequency sub-image may be obtained from the decomposition module 202 , the transformation module 203 , the memory 140 , the input / output device 150 , or an external data storage device via the network 160 .

[0202] At 1102, a composite image may be reconstructed based on the (transformed) low-frequency sub-image and the (transformed) high-frequency sub-image acquired at 1101. Operation 1102 may be performed by the reconstruction module 204. In some embodiments, the reconstructed composite image may correspond to a target image (e.g., an initial image, a region of interest of the initial image, an image of the region of interest, any image generated during image processing, or any combination thereof). In some embodiments, the composite image may be an enhanced target image, a compressed target image, a transformed target image, or any combination thereof. For example, compared to the target image, the composite image may have improved contrast, enhanced details, more distinct edges, or any combination thereof.

[0203] In some embodiments, 1102 may be performed based on one or more reconstruction algorithms. The reconstruction algorithm may include an analytical reconstruction algorithm, an iterative reconstruction algorithm, or any combination thereof. The analytical reconstruction algorithm may include a filtered back projection (FBP) algorithm, a back projection filter (BFP) algorithm, a p-filtered layer graph, or the like. The iterative reconstruction algorithm may include an ordered subset expectation maximization (OSEM) algorithm, a maximum likelihood expectation maximization (MLEM) algorithm, or the like.

[0204] In some embodiments, a composite image can be reconstructed by superimposing the (transformed) low-frequency sub-image and the (transformed) high-frequency sub-image. For example, the grayscale value of an element in the (transformed) low-frequency sub-image can be added to the grayscale value of the same element in the (transformed) high-frequency sub-image to obtain a composite image.

[0205] In some embodiments, the composite image can be reconstructed based on a reconstruction algorithm corresponding to the decomposition algorithm. For example, if a wavelet transform is used to generate a (transformed) low-frequency sub-image and a (transformed) high-frequency sub-image, an inverse wavelet transform can be used to reconstruct the composite image. In some embodiments, the composite image can be further used in a subsequent reconstruction process. For example, the composite image generated together with the sub-image in 1102 can be further used to reconstruct a new composite image. In some embodiments, as Figure 11-B and Figure 11-C As shown, the composite image can be reconstructed based on two or more layers. In some embodiments, the composite image can be sent to the image acquisition module 201, the decomposition module 202, the transformation module 203, the storage 140, the network 160, etc., or any combination thereof.

[0206] Figure 11-B 1 is a flowchart of an exemplary process of reconstructing a low-frequency sub-image of the L'th layer generated by the second decomposition according to some embodiments of the present application. The process may include enhancing the sub-image 1111 and reconstructing the sub-image 1112. It should be understood that according to some embodiments of the present application, Figure 11-BThe process shown can be applied to reconstruct the low-frequency sub-image of the L'th layer generated from the first decomposition.

[0207] In 1111, the high frequency sub-images from the (L'+1)th layer to the (L'+N)th layer generated from the second decomposition may be enhanced. In some embodiments, the high frequency sub-images may be enhanced based on one or more enhancement techniques. The enhancement techniques may include filtering using morphological operators, histogram equalization, noise removal using, for example, Wiener filtering techniques, linear or nonlinear contrast adjustment, median filtering, unsharp mask filtering, contrast limited adaptive histogram equalization (CLAHE), decorrelation stretching, etc., or any combination thereof. In some embodiments, the enhancement techniques may refer to linear / nonlinear enhancement. In some embodiments, linear enhancement may include maximum-minimum contrast techniques, percentage contrast techniques, and segmented contrast techniques, etc., or any combination thereof. Nonlinear enhancement may include histogram equalization, adaptive histogram equalization, homomorphic filtering techniques, unsharp masking, etc., or any combination thereof. In some embodiments, the enhancement techniques may include grayscale value transformation (such as Figure 5-B As shown), weight transformation ( Figure 9-C ), etc., or any combination thereof. In some embodiments, the high-frequency sub-images from the (L'+1)th layer to the (L'+N)th layer generated from the second decomposition can be enhanced by linear / nonlinear enhancement. For example, the high-frequency sub-images from the (L'+1)th layer generated from the second decomposition can be enhanced by a maximum-minimum contrast technique. In some embodiments, different techniques can be used to enhance the high-frequency sub-images from the (L'+1)th layer to the (L'+N)th layer.

[0208] In 1112, the low-frequency sub-image of the L'th layer generated from the second decomposition can be reconstructed. In some embodiments, the low-frequency sub-image of the L'th layer generated from the second decomposition can be reconstructed based on the high-frequency sub-images from the (L'+1)th layer to the (L'+N)th layer enhanced in 1111. In some embodiments, the enhanced high-frequency sub-image of the (L'+N)th layer and the low-frequency sub-image of the (L'+N)th layer can be used to reconstruct the low-frequency sub-image of the (L'+N-1)th layer. The reconstructed low-frequency sub-image of the (L'+N-1)th layer and the enhanced high-frequency sub-image of the (L'+N-1)th layer can be used to reconstruct the low-frequency sub-image of the (L'+N-2)th layer. Therefore, the reconstructed low-frequency sub-image of the (L'+1)th layer and the enhanced high-frequency sub-image of the (L'+1)th layer can be used to reconstruct the low-frequency sub-image of the L'th layer. For example, if L' is equal to 3 and N is equal to 2, the low-frequency sub-image of the fourth layer can be reconstructed based on the low-frequency sub-image of the fifth layer and the enhanced high-frequency sub-image of the fifth layer. Then, the low frequency sub-image of the third layer can be reconstructed based on the reconstructed low frequency sub-image of the fourth layer and the enhanced high frequency sub-image of the fourth layer. In some embodiments, the low frequency sub-image of the L′th layer can be further used to reconstruct a composite image by the reconstruction module 204 . Figure 11-C It should be noted that in some embodiments, in 1111 and / or 1112 , the low-frequency sub-images from the (L′+1)th layer to the (L′+N)th layer may be enhanced before and / or after reconstruction.

[0209] Figure 11-C 1 is a flowchart of an exemplary process for reconstructing a composite image based on the L layer generated by the first decomposition according to some embodiments of the present application. The process may include 1121 for updating the sub-image, one or more operations (e.g., 1122, 1123, etc.) for reconstructing the sub-image, and 1124 for reconstructing the composite image. In some embodiments, the first decomposition and the corresponding reconstruction can improve the detail information of the target image. In some embodiments, the second decomposition and the corresponding reconstruction can improve the edge information of the target image. The reconstruction using the first decomposed sub-image and the second decomposed sub-image can correspondingly improve the detail information and edge information of the target image, and / or enhance the contrast of the target image. In addition, the artifacts of the target image can be reduced by transforming the first decomposition and / or the second decomposition sub-image.

[0210] In 1121, the low-frequency sub-image of the Lth layer generated from the first decomposition may be updated. In some embodiments, the low-frequency sub-image of the Lth layer may be updated based on the low-frequency sub-image of the L'th layer generated from the second decomposition (e.g., Figure 11-BAs shown). In some embodiments, the low-frequency sub-image of the Lth layer generated from the first decomposition can be replaced by the (transformed) low-frequency sub-image of the L'th layer generated from the second decomposition. In some embodiments, L can be equal to L'. In some embodiments, the sub-image generated from the first decomposition and the sub-image generated from the second decomposition can be derived from the same target image. In some embodiments, the low-frequency sub-image of the Lth layer can be updated by transformation. In some embodiments, the updated low-frequency sub-image of the Lth layer can be further transformed. The transformation technology may include grayscale value transformation, weight transformation, linear / non-linear enhancement, upsampling, etc., or any combination thereof. For example, the updated low-frequency sub-image of the Lth layer can be upsampled using bilinear interpolation. In some embodiments, the updated low-frequency sub-image of the Lth layer can be used to reconstruct the low-frequency sub-image of the (L-1)th layer in 1122.

[0211] At 1122, the low-frequency sub-image of the (L-1)th layer generated from the first decomposition may be updated. In some embodiments, the low-frequency sub-image of the (L-1)th layer may be reconstructed based on the updated low-frequency sub-image of the Lth layer generated from the first decomposition and the high-frequency sub-image of the Lth layer. In some embodiments, the updated low-frequency sub-image of the Lth layer may be upsampled by the upsampling unit 203-4. In some embodiments, the updated low-frequency sub-image of the (L-1)th layer may be further transformed. For example, the low-frequency sub-image of the (L-1)th layer may be upsampled by the upsampling unit 203-4.

[0212] At 1123, the low-frequency sub-image of the (L-2)th layer generated from the first decomposition may be updated. In some embodiments, the low-frequency sub-image of the (L-2)th layer may be reconstructed based on the updated low-frequency sub-image of the (L-1)th layer generated from the first decomposition and the high-frequency sub-image of the (L-1)th layer. In some embodiments, the updated low-frequency sub-image of the (L-2)th layer may be further transformed. For example, the low-frequency sub-image of the (L-2)th layer may be upsampled by the upsampling unit 203-4.

[0213] Similarly, the low-frequency sub-image of the first layer generated from the first decomposition can be updated. In some embodiments, the updated low-frequency sub-image of the first layer can be further transformed. A composite image is reconstructed based on the updated low-frequency sub-image of the first layer generated from the first decomposition and the high-frequency sub-image of the first layer.

[0214] It should be noted that one or more operations for updating the low-frequency sub-images of different layers generated from the first decomposition may be added between 1123 and 1124. In some embodiments, one or more high-frequency sub-images of the first layer to the Lth layer may be transformed before or after the updating operation.

[0215] For illustrative purposes, an exemplary process may be described below. It should be noted that the following description is merely an example and should not be construed as the only embodiment. Those skilled in the art will appreciate that, once the basic principles of operation are understood, the flow chart may be modified or altered without departing from the principles. Such modifications and alterations remain within the scope of the present application.

[0216] In some embodiments, the target image can be decomposed into 3 layers by Laplace transform, and the target image can be decomposed into 5 layers by wavelet transform. The low-frequency sub-image of the third layer generated by Laplace transform can be updated by the low-frequency sub-image of the third layer generated by wavelet transform.

[0217] By way of example only, the low-frequency sub-image of the fifth layer generated by the wavelet transform may be transformed using bilinear interpolation; the high-frequency sub-image of the fifth layer generated by the wavelet transform may be transformed using nonlinear enhancement. The low-frequency sub-image of the fourth layer may be reconstructed based on the transformed low-frequency sub-image of the fifth layer and the transformed high-frequency sub-image of the fifth layer. The reconstructed low-frequency sub-image of the fourth layer may be transformed using bilinear interpolation. The high-frequency sub-image of the fourth layer generated by the wavelet transform may be transformed using nonlinear enhancement. The low-frequency sub-image of the third layer may be reconstructed based on the transformed low-frequency sub-image of the fourth layer and the transformed high-frequency sub-image of the fourth layer.

[0218] The low-frequency sub-image of the third layer generated by Laplace transform can be replaced by the reconstructed low-frequency sub-image of the third layer generated by wavelet transform. The updated low-frequency sub-image of the third layer generated by Laplace transform can be further used to update the sub-images of the second and first layers generated by Laplace transform.

[0219] By way of example only, the updated low-frequency sub-image of the third layer generated by Laplace transform may be transformed using bilinear interpolation; and the high-frequency sub-image of the third layer generated by Laplace transform may be transformed using nonlinear enhancement. The transformed low-frequency sub-image of the third layer and the transformed high-frequency sub-image of the third layer may be used to reconstruct the low-frequency sub-image of the second layer. The updated low-frequency sub-image of the second layer may be transformed using bilinear interpolation. The high-frequency sub-image of the second layer generated by Laplace transform may be transformed using nonlinear enhancement. The transformed low-frequency sub-image of the second layer and the transformed high-frequency sub-image of the second layer may be used to reconstruct the low-frequency sub-image of the first layer. The updated low-frequency sub-image of the first layer may be transformed using bilinear interpolation. The high-frequency sub-image of the first layer generated by Laplace transform may be transformed using nonlinear enhancement. The transformed low-frequency sub-image of the first layer and the transformed high-frequency sub-image of the first layer may be used to reconstruct the composite image.

[0220] In some embodiments, the information in the L layer generated by the Laplace transform and the information in the L'+N layer generated by the wavelet transform can be combined. The low-frequency sub-image of the Lth layer generated by the wavelet transform can be reconstructed based on the low-frequency sub-image in the subsequent layer and the enhanced high-frequency sub-image in the subsequent layer. The updated low-frequency sub-image of the Lth layer generated by the Laplace transform and the enhanced high-frequency sub-images of the first to Lth layers can be used to reconstruct the enhanced image. As a result, details and / or edges in the target image can be enhanced.

[0221] In some embodiments, different interpolation algorithms can be used to reconstruct low-frequency sub-images of different layers generated by wavelet transform and / or Laplace transform. As a result, image contrast can be enhanced, and artifacts can be removed from the enhanced image. Thus, image quality can be improved.

[0222] In some embodiments, one or more high-frequency sub-images can be updated based on the weight transformation. In some embodiments, one or more high-frequency sub-images can be enhanced by enhancement unit 203-3 and / or denoised based on a denoising algorithm. For example, a Gaussian filter can be used to denoise the high-frequency sub-images of a layer. The denoised high-frequency sub-images can be used to reconstruct the low-frequency sub-images.

[0223] It should be noted that the above description of image acquisition, image decomposition, image transformation, and image reconstruction is for illustrative purposes only. Those skilled in the art, once they understand the underlying principles of operation, can modify, combine, or alter the processing procedures without departing from the principles. Such modifications, combinations, or alterations remain within the scope of the present application. An exemplary process is described below.

[0224] First, a breast image can be generated from a CT image of the chest. Figure 3-B As shown, a breast region can be extracted from a CT image, a breast edge can be extracted from the CT image, and a breast image can be generated based on the breast region and the breast edge.

[0225] Secondly, the breast image can be decomposed into a low-frequency sub-image and a high-frequency sub-image. The low-frequency sub-image may include information about the breast edge.

[0226] Again, a grayscale value transform may be used to transform the low frequency sub-image of the breast image. In some embodiments, during the process of generating a CT image of the chest, the breast may be compressed. The thickness of the compressed breast may be non-uniform, and the grayscale values in the breast image may be non-uniform. For example, the brightness of elements at the edge of the breast may be darker than the brightness of areas adjacent to the edge of the breast. The grayscale values of elements at the edge of the breast may be close to the grayscale values of the background. Figure 5-B The grayscale value transformation technique shown in can improve the image quality of breast images. Figure 5-B As shown, a characteristic curve can be determined based on the low-frequency sub-image. In some embodiments, the characteristic curve can be divided into N segments. In some embodiments, a transformation curve segment can be determined based on the N divided characteristic curve segments. Then, a transformation curve can be generated based on the transformation curve segments. In some embodiments, the grayscale values of elements in a preset region of the low-frequency sub-image can be updated based on the transformation curve. The preset region may refer to an area where the distance between elements is within a preset value. The grayscale values in the preset region may be close to the grayscale values in the neighborhood.

[0227] Finally, the updated low-frequency sub-image and high-frequency sub-image can be used to reconstruct a synthesized breast image. The grayscale values of elements in the synthesized breast image can be uniform, and the thickness of the breast can be uniform.

[0228] The basic concepts have been described above. It will be apparent to those skilled in the art that the above disclosures are merely illustrative and do not constitute limitations on this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and amendments to this application. Such modifications, improvements, and amendments are suggested in this application and remain within the spirit and scope of the exemplary embodiments of this application.

[0229] At the same time, this application uses specific terms to describe the embodiments of this application. For example, "one embodiment," "an embodiment," and / or "some embodiments" refer to a certain feature, structure, or characteristic related to at least one embodiment of this application. Therefore, it should be emphasized and noted that "one embodiment," "an embodiment," or "an alternative embodiment" mentioned twice or multiple times in different locations in this application does not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this application may be appropriately combined.

[0230] In addition, it will be understood by those skilled in the art that various aspects of the present application can be illustrated and described by a number of patentable categories or situations, including any new and useful process, machine, product or combination of substances, or any new and useful improvements thereto. Accordingly, various aspects of the present application can be implemented entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above hardware or software may be referred to as a "unit," "module," or "system." In addition, various aspects of the present application may be manifested as a computer product located in one or more computer-readable media, which includes computer-readable program code.

[0231] A computer-readable signal medium may include a propagated data signal containing computer program code, for example, in baseband or as part of a carrier wave. The propagated signal may have a variety of manifestations, including electromagnetic, optical, etc., or a suitable combination thereof. A computer-readable signal medium may be any computer-readable medium other than a computer-readable storage medium that can be connected to an instruction execution system, device, or apparatus to communicate, propagate, or transmit a program for use. The program code on the computer-readable signal medium may be propagated via any suitable medium, including radio, cable, fiber optic cable, RF, or similar media, or any combination of the above.

[0232] The computer program code required for the operation of each part of the present application can be written in any one or more programming languages, including object-oriented programming languages such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc., conventional procedural programming languages such as C language, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, dynamic programming languages such as Python, Ruby and Groovy, or other programming languages. The program code can be run entirely on the user's computer, or as a separate software package on the user's computer, or partly on the user's computer and partly on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any network form, such as a local area network (LAN) or a wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as software as a service (SaaS).

[0233] In addition, unless expressly stated in the claims, the order of the processing elements and sequences described in this application, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of this application. Although the above disclosure discusses some of the invention embodiments currently considered useful through various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the essence and scope of the embodiments of this application. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0234] Similarly, it should be noted that in order to simplify the description of the present disclosure and thus facilitate understanding of one or more embodiments of the invention, the foregoing description of the embodiments of the present disclosure sometimes combines multiple features into a single embodiment, figure, or description thereof. However, this disclosure method does not mean that the subject matter of the present disclosure requires more features than those mentioned in the claims. In fact, the features of the patent application scope are fewer than all the features of the individual embodiments disclosed above.

[0235] In some embodiments, the numbers used to describe and claim certain embodiments of the present application, such as the amounts of ingredients, properties such as molecular weight, reaction conditions, etc., should be understood to be in some cases referred to by the terms "about," "approximately," or "substantially." Unless otherwise indicated, "about," "approximately," or "substantially" indicate that the number described allows for a variation of ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may vary depending on the desired characteristics of individual embodiments. In some embodiments, the numerical parameters should be considered to have specified significant digits and adopt a general method of retaining digits. Although the numerical ranges and parameters used to identify the breadth of their ranges in some embodiments of the present application are approximate, in specific embodiments, such numerical values are set as accurately as possible within the feasible range.

[0236] All patents, patent applications, patent application publications, and other materials, such as articles, books, specifications, publications, documents, things, etc., referred to in this application are hereby incorporated by reference in their entirety into this application for all purposes, except any prosecution record related to this document, any document that is inconsistent or conflicting with this document, or any document that has the effect of eventually limiting the broadest scope of the claims related to this document. For example, if there is any inconsistency or conflict between the description, definitions, and / or use of terms associated with any incorporated material and the terminology associated with this document, the description, definitions, and / or use of terms in this document will control.

[0237] Finally, it should be understood that the embodiments described in this application are merely illustrative of the principles of the embodiments of this application. Other variations may also fall within the scope of this application. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this application may be considered consistent with the teachings of this application. Accordingly, the embodiments of this application are not limited to the embodiments explicitly introduced and described in this application.

Claims

1. An image processing method implemented on at least one machine, each of the machines having at least one processor and a memory, the method comprising: Obtaining a target image, the target image including a plurality of elements corresponding to pixels or voxels; Decomposing the target image into at least one layer, the at least one layer including a low-frequency sub-image and a high-frequency sub-image, the low-frequency sub-image including a preset region including a plurality of gray values; Transforming the at least one layer, including transforming the plurality of gray values of the preset region; And Reconstructing the transformed layer into a composite image; Wherein, the transforming the plurality of gray values of the preset region includes: Determining a reference edge in the low-frequency sub-image; Determining a characteristic curve based on the low-frequency sub-image, the characteristic curve representing the relationship between the distance and the gray value corresponding to the distance, wherein the distance refers to the distance between a first element in the low-frequency sub-image and a second element in the reference edge, the first element corresponding to the second element, and the gray value being determined based on the plurality of gray values; Determining a transformation curve based on the characteristic curve, the transformation curve representing the relationship between the gray value before transformation and the gray value after transformation; and Updating the plurality of gray values of the preset region based on the transformation curve.

2. The method according to claim 1, wherein the obtaining the target image includes: Obtaining an initial image; Extracting a region of interest ROI based on the initial image; Extracting the edge of the region of interest based on the initial image; And Determining an image of the region of interest as the target image based on the region of interest and the edge of the region of interest.

3. The method according to claim 1, wherein the determining the transformation curve includes: Dividing the characteristic curve into N characteristic curve segments; Determining N transformation curve segments based on the N characteristic curve segments, wherein one characteristic curve segment corresponds to one transformation curve segment; and Generating the transformation curve based on the N transformation curve segments.

4. The method according to claim 3, wherein the determining the N transformation curve segments includes: For the x-th transformation curve segment of the N transformation curve segments, Calculating the slope of the x-th transformation curve segment based on the gray value of a preset point in the characteristic curve, the gray value of the initial point of the x-th characteristic curve segment, and the gray value of the end point of the x-th characteristic curve segment, the x-th characteristic curve segment corresponding to the x-th transformation curve segment, where x is an integer and 1 ≤ x ≤ N; Determining the gray value of the initial point in the x-th transformation curve segment includes: If x = 1, designating the gray value of the initial point in the x-th characteristic curve segment as the gray value of the initial point in the x-th transformation curve segment; and If 1 < x ≤ N, determining the gray value of the initial point in the x-th transformation curve segment based on the gray value of the initial point of the (x - 1)-th transformation curve segment and the gray value change amount of the (x - 1)-th characteristic curve segment.

5. The method according to claim 3, further comprising: determining a grayscale interval of the characteristic curve, in which at least one grayscale value is to be transformed, the grayscale interval corresponding to a portion of the characteristic curve; as well as The maximum value or the minimum value of the grayscale interval is designated as the grayscale value of a preset point in the characteristic curve.

6. The method of claim 1 , wherein the high-frequency sub-image comprises a plurality of elements, and wherein transforming the at least one layer comprises: generating a weight map of the high-frequency sub-image, wherein the weight map includes a plurality of weights corresponding to the plurality of elements; as well as The high frequency sub-image is updated based on the weight map.

7. The method of claim 6, wherein the high-frequency sub-image comprises first-category elements and second-category elements, and generating the weight map comprises: determining a grayscale interval of the first type of elements in the high-frequency sub-image; determining, based on the grayscale interval of the first-category elements, the grayscale interval of the second-category elements in the high-frequency sub-image; Mapping the grayscale interval of the first type of elements to [0, 1]; Determining weights of the first-category elements based on the mapped grayscale intervals of the first-category elements; Mapping the grayscale interval of the second type element to (1, G], where G is a preset value; Determining weights of the second-category elements based on the mapped grayscale intervals of the second-category elements; as well as The weight graph is generated based on the weights of the first-category elements and the weights of the second-category elements.

8. The method according to claim 7, wherein determining the grayscale interval of the first type of elements comprises: determining an initial grayscale interval of the first type of elements based on a grayscale threshold; Modifying the initial grayscale interval of the first type of elements; as well as Based on the modified grayscale interval of the first-category elements, the initial grayscale interval of the first-category elements is adjusted.

9. An image processing system comprising: An image acquisition module, configured to acquire a target image, wherein the target image includes a plurality of elements corresponding to pixels or voxels; a decomposition module, configured to decompose the target image into at least one layer, wherein the at least one layer includes a low-frequency sub-image and a high-frequency sub-image, wherein the low-frequency sub-image includes a preset region, and the preset region includes a plurality of grayscale values; a transformation module, configured to transform the at least one layer, including transforming the plurality of grayscale values of the preset area; as well as a reconstruction module, configured to reconstruct the transformed layer into a composite image; In order to transform the multiple grayscale values of the preset area, the transformation module is further used to: determining a reference edge in the low-frequency sub-image; determining a characteristic curve based on the low-frequency sub-image, the characteristic curve representing a relationship between a distance and a grayscale value corresponding to the distance, wherein the distance is a distance between a first element in the low-frequency sub-image and a second element in the reference edge, the first element corresponding to the second element, and the grayscale value is determined based on the multiple grayscale values; determining a transformation curve based on the characteristic curve, wherein the transformation curve represents a relationship between a grayscale value before transformation and a grayscale value after transformation; as well as The plurality of grayscale values of the preset area are updated based on the transformation curve.

10. A computer-readable storage medium storing computer instructions, wherein when a computer reads the computer instructions in the storage medium, the computer executes the method according to any one of claims 1 to 8.

Images