A training method and system for image processing models

Abstract

The embodiments of this specification disclose a method and system for training an image processing model. The method includes: obtaining multiple training samples, each of which includes a sample low-energy image; inputting the training samples into the image processing model to determine a base material density image corresponding to the sample low-energy image; and adjusting parameters of the image processing model based on the base material density image, one or more of the labeled base material density image, the sample low-energy image, the sample high-energy image, and sample topological data, with the optimization of a target loss function as the training objective, to obtain a trained image processing model.

Description

Technical Field

[0001] This specification relates to the field of image processing, and in particular to a training method and system for an image processing model. Background Art

[0002] When a CT (Computed Tomography) machine scans, a radiation source emits X-rays toward the inspected part of the target object, and a detector receives the attenuation signal of the X-rays that pass through the inspected part. The computer then reconstructs a tomographic image of the inspected part of the target object. This machine has the advantages of fast scanning time and clear images, and can be used to detect a variety of diseases.

[0003] Dual-energy (DE) CT. Several methods exist for performing dual-energy CT acquisition, such as dual sources, rapid voltage (kVp) switching, and dual-layer detector configurations. Compared to conventional CT, DE CT offers higher detection accuracy and can accurately determine the material information of the scanned object. It is a specialized configuration of spectral CT that utilizes two attenuation values ​​acquired at two different energy spectra to resolve photoelectric and Compton contributions, which include the material's mass attenuation coefficient. Unknown materials can thus be identified based on their photoelectric and Compton contributions. Iodine can be distinguished from, for example, calcium and water due to its photoelectric / Compton properties. Because any two linearly independent sums of two basis functions span the entire attenuation coefficient space, any material can be represented by a linear combination of two other materials (so-called basis materials), such as water and iodine. This offers new applications, such as monochromatic images, material elimination images, effective atomic number images, and electron density images. Performing material decomposition on high-energy, low-energy, or conventionally scanned images to obtain bases or combinations of bases plays a significant role in numerous applications, including automated bone and contrast agent separation in enhanced scans, iodine quantification, qualitative analysis of lesions such as kidney stones, and the generation of pseudo-monoenergetic images and virtual non-enhanced images. Therefore, it is essential to obtain bases with high signal-to-noise ratios. However, existing dual-energy CT imaging methods face several practical limitations. For example, direct material decomposition of dual-energy CT images via matrix inversion can degrade the signal-to-noise ratio of the base material density image.

[0004] Therefore, it is necessary to propose a training method for an image processing model so as to obtain a base material density image with a high signal-to-noise ratio through the trained image processing model. Summary of the Invention

[0005] One aspect of an embodiment of the present specification provides a method for training an image processing model. The method includes: obtaining multiple training samples, wherein each of the multiple training samples includes a sample low-energy image; inputting the training samples into the image processing model to determine a base material density image corresponding to the sample low-energy image; based on the base material density image, one or more of the labeled base material density image, the sample low-energy image, the sample high-energy image, and the sample topology data, and the base material density image, with the optimization of the target loss function as the training objective, adjusting the parameters of the image processing model to obtain a trained image processing model; wherein the sample high-energy image corresponds to the sample low-energy image.

[0006] Another aspect of the embodiments of this specification provides a training system for an image processing model. The system includes: a first acquisition module for acquiring multiple training samples, wherein each of the multiple training samples includes a sample low-energy image; a first determination module for inputting the training samples into the image processing model and determining the base material density image corresponding to the sample low-energy image; a parameter adjustment module for adjusting the parameters of the image processing model based on one or more of the labeled base material density image, the sample low-energy image, the sample high-energy image, and the sample topology data, and the base material density image, with minimizing the target loss function as the training objective, to obtain a trained image processing model; wherein the sample high-energy image corresponds to the sample low-energy image.

[0007] Another aspect of an embodiment of the present specification provides a method for generating a base material density image, the method comprising: obtaining an image to be processed of a target object; inputting the image to be processed into an image processing model trained by the above-mentioned image processing model training method, and determining the base material density image of the image to be processed.

[0008] Another aspect of an embodiment of the present specification provides a system for generating a base material density image, the system comprising: a second acquisition module for acquiring an image to be processed of a target object; a second determination module for inputting the image to be processed into an image processing model trained using the above-mentioned image processing model training method to determine the base material density image of the image to be processed.

[0009] Another aspect of an embodiment of the present specification provides a method for generating a high-energy image, the method comprising: obtaining a low-energy image and topological data of a target object; the topological data comprising low-energy topological data and high-energy topological data, the low-energy topological data and the high-energy topological data corresponding to each other; inputting the low-energy image into an image processing model trained by the training method of the above-mentioned image processing model, and determining a base material density image of the low-energy image; determining the difference in topological data based on the low-energy topological data and the high-energy topological data; and determining a high-energy image corresponding to the low-energy image based on the difference in the low-energy image, the base material density image and the topological data.

[0010] Another aspect of an embodiment of the present specification provides a system for generating a high-energy image, the system comprising: a third acquisition module, for acquiring a low-energy image and topological data of a target object; the topological data comprises low-energy topological data and high-energy topological data, and the low-energy topological data and the high-energy topological data correspond to each other; a third determination module, for inputting the low-energy image into an image processing model trained by the training method of the above-mentioned image processing model, and determining a base material density image of the low-energy image; a fourth determination module, for determining the difference in topological data based on the low-energy topological data and the high-energy topological data; and a fifth determination module, for determining the high-energy image corresponding to the low-energy image based on the difference in the low-energy image, the base material density image and the topological data.

[0011] Another aspect of an embodiment of this specification provides a training device for an image processing model including at least one storage medium and at least one processor, wherein the at least one storage medium is used to store computer instructions; and the at least one processor is used to execute the computer instructions to implement a training method for an image processing model.

[0012] Another aspect of the embodiments of this specification provides a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the training method of the image processing model.

[0013] In the embodiments of the present specification, the image processing model is trained using sample low-energy images, and the trained image processing model can realize the material decomposition function through low-energy images; at the same time, the hardware implementation of the embodiments of the present specification is simple, and the single-energy CT image obtained by scanning the target object with a lower dose of rays can obtain the base material density image, which reduces the radiation dose to the target object, and the obtained base material density image can have a higher signal-to-noise ratio. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] This specification will be further described in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, like numbers represent like structures, wherein:

[0015] Figure 1 is a schematic diagram of an exemplary application scenario of an image processing system according to some embodiments of this specification;

[0016] Figure 2 is an exemplary flow chart of a method for training an image processing model according to some embodiments of this specification;

[0017] Figure 3 is an exemplary flow chart for determining a base material density image loss function according to some embodiments of this specification;

[0018] Figure 4 is an exemplary flow chart of determining a first high-energy image loss function according to some embodiments of this specification;

[0019] Figure 5 is an exemplary flow chart of determining a second high-energy image loss function according to some embodiments of this specification;

[0020] Figure 6 is an exemplary flow chart for determining a topological high-energy image according to some embodiments of this specification;

[0021] Figure 7 is an exemplary module diagram of a training system for an image processing model according to some embodiments of this specification;

[0022] Figure 8 is an exemplary module diagram of a system for generating a base material density image according to some embodiments of this specification;

[0023] Figure 9 is an exemplary module diagram of a system for generating high-energy images according to some embodiments of the present specification. DETAILED DESCRIPTION

[0024] To more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly describes the drawings required for describing the embodiments. Obviously, the drawings described below are merely examples or embodiments of this specification. Those skilled in the art can apply this specification to other similar scenarios based on these drawings without inventive effort. Unless otherwise apparent from the context or otherwise noted, the same reference numerals in the figures represent the same structure or operation.

[0025] It should be understood that the terms "system," "device," "unit," and / or "module" used herein are a method for distinguishing different components, elements, parts, portions, or assemblies at different levels. However, other terms may be substituted for the terms if they can achieve the same purpose.

[0026] As used in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not refer to the singular but also include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0027] Flowcharts are used throughout this specification to illustrate the operations performed by systems according to embodiments of this specification. It should be understood that preceding or following operations do not necessarily need to be performed in exact order. Instead, the steps may be processed in reverse order or simultaneously. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0028] Computed tomography (CT) is currently widely used in clinical diagnosis. Dual-energy CT technology has developed rapidly and has gradually become one of the commonly used examination methods in clinical diagnosis.

[0029] Currently, one approach to implementing dual-energy CT imaging in clinical practice is through hardware design, such as rapid kVp switching, dual-layer detector technology, dual-source CT imaging, and multiple scans. Direct matrix inversion in the image domain or iterative material decomposition is then performed to obtain the base material density image. Rapid kVp switching utilizes only a single X-ray source, allowing a single tube to rapidly switch between high and low voltages to acquire dual-energy data. However, this typically results in longer scan times, poor discrimination between energy levels, and minimal difference between the generated high-energy and low-energy images. Furthermore, an additional scan is required. Dual-source CT imaging utilizes two X-ray sources, each emitting X-rays at two different energy levels. These two sources are angled relative to their rotational centers, allowing them to emit X-rays simultaneously. The scanning ranges of the low-energy and high-energy rays differ significantly at the same time. While dual-source CT imaging offers good energy level discrimination, its hardware implementation is complex. Furthermore, due to the angular offset between the two sources, the generated dual-energy images are prone to offset or distortion in areas of intense motion, potentially resulting in artifacts in the material decomposition images. In addition, the dual-energy CT technologies exemplified above all have a common problem: the radiation dose to patients is high, and the speed of directly iterative material decomposition of dual-energy CT images is also slow, making it difficult to meet clinical needs.

[0030] This specification discloses an image processing model training method that incorporates machine learning techniques to train the model. The trained image processing model can be used to perform material decomposition using single-energy CT images. The following describes the technical solutions disclosed in this specification in detail, with reference to the accompanying figures.

[0031] Figure 1 It is a schematic diagram of an exemplary application scenario of the image processing system shown in some embodiments of this specification.

[0032] In some embodiments, the image processing system 100 can be used to obtain a corresponding base material density image based on a single-energy CT image. For example, the image processing system 100 can use an image processing model to process a single-energy CT image (e.g., a low-energy CT image) to obtain a corresponding base material density image.

[0033] In some embodiments, the image processing model can be trained in other systems (e.g., a training system for the image processing model, not shown), or can be trained using the image processing system 100. For example, the image processing system 100 can obtain multiple training samples, wherein each of the multiple training samples includes a sample low-energy image; the sample high-energy image corresponds to the sample low-energy image; the image processing system 100 can input the training samples into the image processing model to determine the base material density image corresponding to the sample low-energy image; the image processing system 100 can adjust the parameters of the image processing model based on one or more of the labeled base material density image, the sample low-energy image, the sample high-energy image, and the sample topology data, and the base material density image, with the optimization of the target loss function as the training objective, to obtain a trained image processing model. The sample high-energy image corresponds to the sample low-energy image.

[0034] like Figure 1 As shown, the image processing system 100 may include an imaging device 110 , a network 120 , a terminal 130 , a processing device 140 , and a storage device 150 .

[0035] The imaging device 110 can be used to image the target object to generate an image. The imaging device 110 can be a medical imaging device (e.g., CT (Computed Tomography). In some embodiments, the imaging device 110 can include a gantry 111, a detector 112, a scanning area 113, and a scanning bed 114. The target object can be placed on the scanning bed 114 to be scanned. The gantry 111 can support the detector 112. In some embodiments, the detector 112 can include one or more detector units. The detector unit can be and / or include a single-row detector and / or a multi-row detector. The detector unit can include a scintillation detector (e.g., a cesium iodide detector) and other detectors. In some embodiments, the gantry 111 can rotate. For example, in a CT imaging device, the gantry 111 can rotate clockwise or clockwise around the gantry rotation axis. Rotates counterclockwise. In some embodiments, the imaging device 110 may further include a radiation scanning source, which may rotate together with the gantry 111. The radiation scanning source may emit a radiation beam (e.g., X-rays) to a target object, which is detected by the detector 112 after being attenuated by the target object, thereby generating an image signal. In some embodiments, the scanning bed 114 may be movably disposed in front of the machine and parallel to the ground. The scanning bed 114 may be moved to enter and exit the scanning area 113. The scanning bed 114 may also be moved in a vertical direction to adjust the distance between the target object on the scanning bed and the detector 112 (or the scanning center) when entering the scanning area 113, so as to scan the target object within the scanning range.

[0036] The processing device 140 can process data and / or information obtained from the imaging device 110, the terminal 130, and / or the storage device 150. For example, the processing device 140 can process the image information generated by the detector 112 to obtain a CT image. For another example, the processing device 140 can process the CT image to obtain a base material density image corresponding to the CT image. In some embodiments, the processing device 140 can be a single server or a server group. The server group can be centralized or distributed. In some embodiments, the processing device 140 can be local or remote. For example, the processing device 140 can access information and / or data from the imaging device 110, the terminal 130, and / or the storage device 150 via the network 120. For another example, the processing device 140 can directly connect to the imaging device 110, the terminal 130, and / or the storage device 150 to access information and / or data. In some embodiments, the processing device 140 can be implemented on a cloud platform. For example, the cloud platform can include one or a combination of private clouds, public clouds, hybrid clouds, community clouds, distributed clouds, cross-clouds, and multiple clouds.

[0037] Terminal 130 may include a mobile device 131, a tablet computer 132, a laptop computer 133, or the like, or any combination thereof. In some embodiments, terminal 130 may interact with other components of image processing system 100 via a network. For example, terminal 130 may send one or more control instructions to imaging device 110 to control imaging device 110 to scan a target object according to the instructions. For another example, terminal 130 may also receive a base material density image determined by processing device 140 and display the base material density image for operator analysis and confirmation. In some embodiments, mobile device 131 may include a smart home device, a wearable device, a mobile device, a virtual reality device, an augmented reality device, or the like, or any combination thereof. In some embodiments, smart home devices may include smart lighting devices, smart appliance control devices, smart monitoring devices, smart televisions, smart radiography machines, intercoms, or the like, or any combination thereof. In some embodiments, wearable devices may include bracelets, shoes and socks, glasses, helmets, watches, clothing, backpacks, smart accessories, or the like, or any combination thereof. In some embodiments, mobile devices may include mobile phones, personal digital assistants (PDAs), gaming devices, navigation devices, point-of-sale devices, laptop computers, tablet computers, desktop computers, or the like, or any combination thereof. In some embodiments, the virtual reality device and / or augmented reality device may include a virtual reality helmet, virtual reality glasses, a virtual reality patch, an augmented reality helmet, augmented reality glasses, an augmented reality patch, etc. or any combination thereof. For example, the virtual reality device and / or augmented reality device may include Google Glass TM 、Oculus Rift TM , HoloLens TM or Gear VR TM In some embodiments, the terminal 130 may be part of the processing device 140. In some embodiments, the terminal 130 may be integrated with the processing device 140 to serve as an operating console for the imaging device 110. For example, a user / operator of the image processing system 100 (e.g., a doctor) may control the operation of the imaging device 110 through the operating console, such as scanning a target object, controlling the movement of the scanning bed 114, training an image processing model, and obtaining a base material density image using the image processing model.

[0038] The storage device 150 can store data (e.g., scan data of a target object), instructions, and / or any other information. In some embodiments, the storage device 150 can store data obtained from the imaging device 110, the terminal 130, and / or the processing device 140. For example, the storage device 150 can store treatment plans, scan data of a target object, etc. obtained from the imaging device 110. In some embodiments, the storage device 150 can store data and / or instructions that can be executed or used by the processing device 140 to perform the exemplary methods described herein. In some embodiments, the storage device 150 can include one or a combination of mass storage, removable memory, volatile read-write memory, read-only memory (ROM), etc. Mass storage can include magnetic disks, optical disks, solid-state drives, removable storage, etc. Removable memory can include flash drives, floppy disks, optical disks, memory cards, ZIP disks, magnetic tape, etc. Volatile read-write memory can include random access memory (RAM). RAM may include dynamic random access memory (DRAM), double data rate synchronous dynamic random access memory (DDR-SDRAM), static random access memory (SRAM), thyristor random access memory (T-RAM), zero capacitance random access memory (Z-RAM), etc. ROM may include mask read-only memory (MROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM), digital versatile disc, etc. In some embodiments, the storage device 150 may be implemented by the cloud platform described in this application. For example, the cloud platform may include one or a combination of private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, cross-cloud, multi-cloud, etc.

[0039] In some embodiments, the storage device 150 can be connected to the network 120 to enable communication with one or more components in the image processing system 100 (e.g., the processing device 140, the terminal 130, etc.). One or more components in the image processing system 100 can read data or instructions from the storage device 150 via the network 120. In some embodiments, the storage device 150 can be part of the processing device 140 or independent and directly or indirectly connected to the processing device.

[0040] The network 120 may include any suitable network capable of facilitating information and / or data exchange within the image processing system 100. In some embodiments, one or more components of the image processing system 100 (e.g., the imaging device 110, the terminal 130, the processing device 140, the storage device 150, etc.) may exchange information and / or data with one or more components of the image processing system 100 via the network 120. For example, the processing device 140 may obtain scan data from the imaging device 110 via the network 120. The network 120 may include one or a combination of a public network (e.g., the Internet), a private network (e.g., a local area network (LAN), a wide area network (WAN), etc.), a wired network (e.g., an Ethernet network), a wireless network (e.g., an 802.11 network, a wireless Wi-Fi network, etc.), a cellular network (e.g., a Long Term Evolution (LTE) network), a frame relay network, a virtual private network (VPN), a satellite network, a telephone network, a router, a hub, a server computer, and the like. For example, the network 120 may include a wired network, an optical fiber network, a telecommunication network, a local area network, a wireless local area network (WLAN), a metropolitan area network (MAN), a public switched telephone network (PSTN), a Bluetooth TM Network, ZigBee TM In some embodiments, the network 120 may include one or more network access points. For example, the network 120 may include wired and / or wireless network access points, such as base stations and / or Internet exchange points, through which one or more components of the image processing system 100 may connect to the network 120 to exchange data and / or information.

[0041] Figure 2 200 is an exemplary flow chart of a method for training an image processing model according to some embodiments of this specification. In some embodiments, process 200 may be executed by a processing device (e.g., processing device 140). For example, process 200 may be stored in a storage device (e.g., a built-in storage unit of the processing device or an external storage device) in the form of a program or instruction. When executed, the program or instruction may implement process 200. Process 200 may include the following operations.

[0042] Step 202 , obtaining multiple training samples. In some embodiments, step 202 may be performed by the first obtaining module 710 .

[0043] The training samples include low-energy image data used to train the image processing model.

[0044] In some embodiments, each of the plurality of training samples may include a sample low-energy image. A low-energy image refers to an image obtained by scanning and imaging a target object using low-dose radiation. The target object may include a patient or other medical experimental subject (e.g., experimental mice or other animals). The target object may also be a part of a patient or other medical experimental subject, including organs and / or tissues, such as the heart, lungs, ribs, or abdominal cavity.

[0045] In some embodiments, the processing device may obtain the plurality of training samples by reading from a database, a storage device, or an imaging device.

[0046] Step 204 : Input the training sample into the image processing model to determine the base material density image corresponding to the sample low-energy image. In some embodiments, step 204 may be performed by the first determination module 720 .

[0047] In some embodiments, the processing device may input the training sample (sample low-energy image) into an image processing model, and the image processing model may output a base material density image corresponding to the sample low-energy image.

[0048] The essence of CT imaging is attenuation coefficient imaging. The principle is that X-rays are penetrating and can penetrate objects (e.g., living organisms or objects). Different tissues in an object have different absorption and transmittance rates for X-rays. When X-rays pass through an object, they are attenuated differently in different parts. By measuring the attenuated rays, data from different parts of the object can be obtained. After inputting the acquired data into an electronic computer for processing, a cross-sectional or three-dimensional image of the examined part of the object can be reconstructed. The linear attenuation coefficient of any object can be expressed as a linear combination of selected materials. Since the CT image is linearly related to the linear attenuation coefficient, the CT image of any object can be expressed as a linear combination of selected materials. For example, a CT image can be expressed using Equation (1).

[0049] I=M·A (1)

[0050] Where I represents a CT image with a dimension of N × 1, where N is the total number of pixels in a CT image; M is the base material density image with a dimension of N × m, where m is the type of decomposed material; and A is the material decomposition matrix with a dimension of m × 1.

[0051] The base material density image output by the image processing model can be used to perform constraint training on the image processing model, and can also be used to obtain a high-energy image corresponding to the low-energy image. The relevant description of obtaining the high-energy image can be found in other parts of this specification, such as step 206 and Figure 6 The relevant instructions will not be repeated here.

[0052] In some embodiments, the image processing model may include a U-net model, a V-net model, etc. based on deep learning.

[0053] Step 206: Based on the label base material density image, the sample low-energy image, the sample high-energy image, the sample topology data, and the base material density image, the parameters of the image processing model are adjusted with the optimization of the target loss function as the training objective to obtain a trained image processing model. In some embodiments, step 206 may be performed by parameter adjustment module 730.

[0054] The labeled base material density image refers to a known base material density image corresponding to the sample low-energy image. Each training sample can have a corresponding labeled base material density image.

[0055] In some embodiments, the label-based material density image can be obtained by inversely decomposing the high-energy image matrix corresponding to the dual-energy CT image (for example, dual-basis decomposition, three-basis decomposition, or multi-basis material decomposition, etc.), or performing iterative material decomposition calculations on the conventional dose dual-energy CT image.

[0056] In some embodiments, the sample high-energy image corresponds to the sample low-energy image. Correspondence means that the sample high-energy image and the sample low-energy image are from the same target object, and the scanning angles when acquiring the sample high-energy image and the sample low-energy image can be different. A high-energy image refers to an image obtained by scanning and imaging the target object using high-energy-level radiation, and a low-energy image refers to an image obtained by scanning and imaging the target object using radiation of a lower energy level than the high-energy-level radiation. For example, the sample low-energy image can be obtained by scanning the target object at a first radiation dose, and the sample high-energy image can be obtained by scanning the target object at a second radiation dose; wherein the first radiation dose is lower than the second radiation dose.

[0057] Topological data refers to scanning data obtained by scanning a target object at a certain angle (which can be any angle, for example, a cross-section, a coronal plane, a sagittal plane, etc.). Sample topological data may include sample low-energy topological data and sample high-energy topological data. Sample low-energy topological data refers to topological data obtained by scanning with rays of lower energy levels, and sample high-energy topological data refers to topological data obtained by scanning with rays of higher energy levels. Among them, the sample low-energy topological data in a single training sample corresponds to the sample high-energy topological data, for example, obtained by scanning the same target object. In some embodiments, the scanning angles of the sample low-energy topological data and the sample high-energy topological data are the same. In some embodiments, the sample low-energy topological data and the sample high-energy topological data can be obtained at the same time. The scanning angle of the sample topological data may be different from the scanning angles of the sample low-energy image and the sample high-energy image, so that more information can be contained in the training sample data.

[0058] In some embodiments, the processing device can input the training samples into the image processing model, and the image processing model outputs the prediction result (i.e., the base material density image), and constructs a loss function based on one or more of the labeled base material density image, the sample low-energy image, the sample high-energy image, and the sample topology data to constrain the loss function. By continuously adjusting the parameters of the image processing model, the loss function value corresponding to each training sample is minimized, so that the prediction result of the final image processing model can be more accurate. When the loss function value meets the requirements (for example, is less than a preset value, the loss function value converges) or the iteration reaches a preset number of times, a trained image processing model can be obtained.

[0059] In some embodiments, the target loss function may be any one or a combination of a base material density image loss function, a low-energy image loss function, a first high-energy image loss function, and a second high-energy image loss function. For example, the target loss function may be a combination of the base material density image loss function and the first high-energy image loss function, a combination of the base material density image loss function and the second high-energy image loss function, or a combination of the base material density image loss function, the first high-energy image loss function, and the second high-energy image loss function.

[0060] In some embodiments, a base material density image loss function may be determined based on at least the label base material density image. For example, a base material density image loss function may be constructed based on the base material density image predicted by the image processing model and the label base material density image. Exemplarily, the base material density image loss function may be as shown in Equation (2).

[0061]

[0062] Among them, L Maerial represents the base material density image loss function; I Low is the sample low-energy image, F(I Low ) is the base material density image predicted by the image processing model; Material is the labeled base material density image; N is the total number of training samples used for model training.

[0063] In this embodiment, a loss function is constructed directly based on the labeled base material density image to constrain the trained image processing model, making the model's predictions close to the actual labeled base material density image obtained using dual-energy CT images. Inputting low-energy images into the trained image processing model yields a base material density image, enabling the use of single-energy CT images to perform material decomposition. Because the base material density image can be obtained using only single-energy CT images, the scan time is shortened and the radiation dose to the target object is lower, compared to the multiple scans required to obtain the base material density image using dual-energy CT images.

[0064] In some embodiments, the low-energy image loss function can be determined based on at least the sample low-energy image. For example, the low-energy image loss function can be determined based on the sample low-energy image and the predicted low-energy image. In some embodiments, the predicted low-energy image can be calculated based on the base material density image and the sample low-energy image. For example, the base material density image can be multiplied by the material decomposition matrix corresponding to the sample low-energy image to obtain the predicted low-energy image. Among them, the material decomposition matrix of the sample low-energy image can be obtained by solving the sample low-energy image and the base material density image in a least squares fitting manner. For more details, please refer to Figure 4 The relevant description will not be repeated here.

[0065] Exemplarily, the low-energy image loss function can be expressed as Equation (3).

[0066]

[0067] in, To predict low energy images; I Low is the sample low-energy image; N is the total number of training samples. The value of the low-energy image loss function can reflect the difference between the predicted low-energy image and the sample low-energy image. The predicted low-energy image is obtained based on the base material density image, and thus can indirectly reflect the accuracy of the model's prediction results.

[0068] In some embodiments, the first high-energy image loss function can be determined based on at least the sample high-energy image. For example, the first high-energy image loss function can be constructed based on the sample high-energy image and the predicted high-energy image. In some embodiments, the predicted high-energy image can be calculated based on the base material density image and the sample high-energy image. For example, the base material density image can be multiplied by the material decomposition matrix corresponding to the sample high-energy image to obtain the predicted high-energy image. The material decomposition matrix of the sample high-energy image can be obtained by solving the sample high-energy image and the base material density image in a least squares fitting manner. For more details, see Figure 4 The relevant description will not be repeated here.

[0069] Exemplarily, the first high-energy image loss function L High-1 It can be shown as equation (4).

[0070]

[0071] in, To predict high energy images; I High is the sample high-energy image; N is the total number of training samples.

[0072] In this embodiment, the predicted high-energy image is obtained by using the predicted base material density image and the sample high-energy image, and then the image processing model is constrained by constructing a loss function based on the predicted high-energy image and the sample high-energy image. Since the predicted high-energy image is obtained based on the predicted base material density image, the loss function constructed using the predicted high-energy image and the sample high-energy image can reflect the accuracy of the predicted base material density image to a certain extent. The trained image processing model can achieve the purpose of obtaining the base material density image using a single-energy low-energy image, and the single-energy CT image can be used to realize the material decomposition function.

[0073] In some embodiments, the second high-energy image loss function may be determined based on at least the sample high-energy image and the sample topology data. For example, the second high-energy image loss function may be constructed based on the sample high-energy image and a topological high-energy image obtained based on the sample topology data.

[0074] Details on determining topological high-energy images based on sample topological data can be found in Figure 6 The relevant instructions will not be repeated here.

[0075] In some embodiments, the second high-energy image loss function L High-2 It can be shown as equation (5).

[0076]

[0077] in, is a topological high-energy image; I High is the sample high-energy image; N is the total number of training samples.

[0078] In this embodiment, the model is constrained by obtaining a topological image through the predicted base material density image and sample topological data, and then constructing a loss function based on the topological high-energy image and the sample high-energy image. Since the topological high-energy image is obtained based on the predicted base material density image and the sample topological data, the loss function constructed using the topological high-energy image and the sample high-energy image can also reflect the accuracy of the predicted base material density image to a certain extent.

[0079] Exemplarily, the target loss function of the partial loss function combination is exemplarily shown in the following embodiments. For example, the combination of the base material density image loss function, the low-energy image loss function and the first high-energy image loss function can be as shown in Equation (6).

[0080] L1=λ·L Maerial +L Low +L High-1 (6)

[0081] Where L1 represents the combined loss function of the base material density image loss function, the low-energy image loss function, and the first high-energy image loss function; λ is the weight balancing factor, and λ is a constant; L Maerial is the base material density image loss function, L Low represents the low-energy image loss function; L High-1 represents the first high-energy image loss function.

[0082] In some embodiments, the target loss function of the combination of the base material density image loss function, the low-energy image loss function, the first high-energy image loss function, and the second high-energy image loss function can be as shown in Equation (7).

[0083] L2=λ·L Maerial +L Low +L High-1 +L High-2 (7)

[0084] Where L2 represents the combined loss function of the base material density image loss, the low-energy image loss function, the first high-energy image loss function, and the second high-energy image loss function; λ is the weight balancing factor, and λ is a constant; L Maerial represents the basis material density image loss function, L Low represents the low-energy image loss function; L High-1 represents the first high-energy image loss function, L High-2 Represents the loss function of the second high-energy image. For a more detailed explanation of the loss functions of each part, please refer to the relevant description above and will not be repeated here.

[0085] Preferably, in some embodiments, the image processing model can be constrained by the combined loss function exemplified by equation (7). The target loss function shown in equation (7) has more constraints, and the trained model can have stronger robustness, so that the prediction results can be better.

[0086] The model is constrained by the loss function shown in the example above. Once the loss function value meets preset conditions (e.g., convergence, less than a preset threshold, etc.) or reaches a preset number of iterations, a trained image processing model is obtained. The trained image processing model can be used to obtain a base material density image.

[0087] In some embodiments, the processing device may obtain a base material density image using a trained image processing model through the method described in the following embodiments.

[0088] In some embodiments, the processing device may acquire an image of the target object to be processed, which may be a low-energy CT image obtained by low-dose radiation scanning.

[0089] In some embodiments, the processing device may obtain the image to be processed of the target object by reading from an imaging device, a database, a storage device, calling a data interface, and the like.

[0090] In some embodiments, the processing device may input the image to be processed into an image processing model trained according to the methods described in the embodiments of this specification to determine the base material density image of the image to be processed. Specifically, the processing device may input the image to be processed of the target object into the trained image processing model. After the image processing model processes the image to be processed, it outputs the base material density image of the image to be processed.

[0091] In some embodiments, the processing device may also generate a high-energy image based on the base material density image obtained by the image processing model using the method described in the following embodiments.

[0092] In some embodiments, the processing device may acquire a low-energy image and topological data of the target object; the topological data includes low-energy topological data and high-energy topological data, and the low-energy topological data corresponds to the high-energy topological data.

[0093] For the description of low-energy images and topological data, please refer to the relevant description in steps 202-206, which will not be repeated here. In some embodiments, the processing device can scan the target object using a scanning device to obtain scanning data, reconstruct a low-energy image based on the scanning data, and obtain topological data from the scanning data. For example, the processing device can use the scanning device to perform a low-energy scan to obtain low-energy topological data and a high-energy scan to obtain high-energy topological data, or it can extract low-energy topological data from the scanning data corresponding to the low-energy image and perform a high-energy scan to obtain high-energy topological data. In some embodiments, the processing device can also read the low-energy image and topological data from a storage device.

[0094] The processing device can input the low-energy image into the image processing model trained by the image processing model training method described in the embodiments of this specification to determine the base material density image of the low-energy image.

[0095] The processing device may determine a topological data difference based on the low-energy topological data and the high-energy topological data.

[0096] For the determination of topological data differences, please refer to this manual Figure 6 The relevant instructions will not be repeated here.

[0097] A high-energy image corresponding to the low-energy image is determined based on the low-energy image, the base material density image, and the topological data difference.

[0098] In some embodiments, the processing device can determine material density data based on the base material density image; determine topological data differences based on the low-energy topological data and the high-energy topological data; determine material decomposition matrix differences based on the material density data and the topological data differences; determine image differences based on the material decomposition matrix differences and the base material density image; and determine the high-energy image based on the low-energy image and the image differences.

[0099] The process of determining the high-energy image corresponding to the low-energy image is similar to the process of determining the topological high-energy image described in this specification. The difference lies in the different data used. Therefore, the specific process of determining the high-energy image corresponding to the low-energy image can be found in Figure 6 The relevant instructions will not be repeated here.

[0100] In an embodiment of the present specification, an image processing model is trained using training sample data. The model can be constrained using a loss function constructed based on one or a combination of labeled base material density images, sample low-energy images, sample high-energy images, and sample topology data. The trained model can then implement material decomposition using single-energy CT images. By constraining the model using a combined loss function, the addition of more constraints can improve the robustness of the model, resulting in more accurate predictions from the trained model. Furthermore, since base material density images can be obtained using only single-energy CT images, single-energy CT images offer shorter scan times and lower radiation doses compared to dual-energy CT images, which require multiple scans of the target object. Furthermore, processing images using the trained image processing model avoids significant degradation of the signal-to-noise ratio of the base material density image and reduces the time required to obtain the base material density image, compared to directly obtaining the base material density image through matrix inversion or iterative material decomposition.

[0101] Figure 3 300 is an exemplary flow chart for determining a base material density image loss function according to some embodiments of this specification. In some embodiments, process 300 may be executed by a processing device. For example, process 300 may be stored in a storage device (such as a built-in storage unit of the processing device or an external storage device) in the form of a program or instruction. When the program or instruction is executed, process 300 may be implemented. Figure 3 As shown, process 300 may include the following operations.

[0102] Step 302: Process the sample low-energy image through the image processing model to obtain the base material density image.

[0103] In some embodiments, the processing device may input the sample low-energy image into the image processing model, and the image processing model processes the sample low-energy image and then outputs the base material density image.

[0104] Step 304: Determine the base material density image loss function based on the base material density image and the label base material density image.

[0105] In some embodiments, the determined basis material density image loss function is as shown in Equation (2).

[0106] In some embodiments, the processing device may determine the difference between the base material density image and the label base material density image based on the base material density image loss function (i.e., equation (2)) illustrated in step 204. For example, the base material density image and the label base material density image are substituted into equation (2) for calculation to obtain the value of the base material density image loss function. The magnitude of the value of the base material density image loss function can reflect the difference between the base material density image and the label base material density image. By minimizing the loss function, the difference can be reduced, making the prediction result of the model more accurate.

[0107] Figure 4 This is an exemplary flow chart for determining a low-energy image loss function and a first high-energy image loss function according to some embodiments of this specification. In some embodiments, process 400 can be executed by a processing device. For example, process 400 can be stored in a storage device (such as a built-in storage unit of the processing device or an external storage device) in the form of a program or instruction. When the program or instruction is executed, process 400 can be implemented. Figure 4 As shown, process 400 may include the following operations.

[0108] Step 402: Process the sample low-energy image through the image processing model to obtain the base material density image.

[0109] The process of obtaining the density image of the base material is similar to Figure 3 The process is the same as described in the example, and more details can be found in Figure 3 The relevant description will not be repeated here.

[0110] Step 404: Determine the predicted low-energy image based on the base material density image.

[0111] The predicted low-energy image refers to a low-energy image calculated based on the prediction result of the base material density image of the image processing model (ie, the base material density image).

[0112] In some embodiments, the predicted low-energy image may be determined based on the following equation (8).

[0113]

[0114] in, Represents the predicted low-energy image; M is the predicted base material density image; A low is the material decomposition matrix of the sample low energy image. low Multiply them together to get the predicted low-energy image

[0115] A low The sample low-energy image and base material density image can be solved by the least square fitting method.

[0116] Step 406 : Determine a low-energy image loss function based on the predicted low-energy image and the sample low-energy image.

[0117] In some embodiments, the low-energy image loss function can be the low-energy image loss function exemplified by equation (3) in step 204 above. The predicted low-energy image and the sample low-energy image are substituted into equation (3) to calculate the value of the low-energy image loss function.

[0118] Step 408: Determine a predicted high-energy image based on the base material density image.

[0119] The predicted high-energy image refers to a high-energy image calculated based on the prediction result of the base material density image (ie, the base material density image) of the image processing model.

[0120] In some embodiments, the predicted low-energy image may be determined based on the following equation (9).

[0121]

[0122] in, Represents the predicted high energy image; M is the predicted base material density image; A High is the material decomposition matrix of the sample high energy image. High Multiply them together to get the predicted high-energy image

[0123] In some embodiments, A high The sample high-energy image and base material density image can be solved by the least square fitting method.

[0124] Step 410: Determine a first high-energy image loss function based on the predicted high-energy image and the sample high-energy image.

[0125] In some embodiments, the first image loss function can be as shown in equation (4) as illustrated in step 204 above. The predicted high-energy image and the sample high-energy image are substituted into equation (4) to calculate the value of the first high-energy image loss function.

[0126] Figure 5 This is an exemplary flow chart for determining the second high-energy image loss function according to some embodiments of this specification. In some embodiments, process 500 can be executed by a processing device. For example, process 500 can be stored in a storage device (such as a built-in storage unit of the processing device or an external storage device) in the form of a program or instruction. When the program or instruction is executed, process 500 can be implemented. Figure 5 As shown, process 500 may include the following operations.

[0127] Step 502: Determine a topological high-energy image based on the base material density image, the sample low-energy image, and the sample topology data.

[0128] A topological high-energy image refers to a high-energy image calculated based on the prediction results of the image processing model (i.e., the base material density image), the sample low-energy image, and the sample topological data.

[0129] The sample topological data includes sample low-energy topological data and sample high-energy topological data.

[0130] Details on determining the topological high-energy image can be found in Figure 6 The related descriptions will not be repeated here.

[0131] Step 504 : Determine the second high-energy image loss function based on the sample high-energy image and the topological high-energy image.

[0132] In some embodiments, the determined second high-energy image loss function may be as shown in equation (5) of step 204 .

[0133] In some embodiments, the processing device may substitute the topological high-energy image and the sample high-energy image into the second high-energy image loss function to calculate the value of the second high-energy image loss function.

[0134] Figure 6 This is an exemplary flow chart for determining a topological high-energy image according to some embodiments of this specification. In some embodiments, process 600 can be executed by a processing device. For example, process 600 can be stored in a storage device (such as a built-in storage unit of the processing device or an external storage device) in the form of a program or instruction, and when the program or instruction is executed, process 600 can be implemented. Figure 6 As shown, process 600 may include the following operations.

[0135] Step 602: Determine sample material density data based on the base material density image.

[0136] The material density data is the data obtained by forward projecting the base material density image. The material density data can be used to determine the difference in the decomposition matrix between the high-energy image and the low-energy image.

[0137] In some embodiments, the processing device may forward project the base material density image to obtain the sample material density data.

[0138] Step 604: Determine the sample topology data difference based on the sample low-energy topology data and the sample high-energy topology data.

[0139] In some embodiments, the processing device may subtract the sample low-energy topological data from the sample high-energy topological data to obtain the sample topological data difference.

[0140] Step 606: Determine the material decomposition matrix difference based on the difference between the sample material density data and the sample topology data.

[0141] In some embodiments, for dual-energy CT images, it can be expressed by the following equations (10) and (11), respectively.

[0142] I Low =M·A Low (10)

[0143] I High =M·A High (11)

[0144] Among them, I Low Represents low-energy image, I High Represents high-energy image, A Low is the material decomposition matrix of the low-energy image, A High is the material decomposition matrix of the high-energy image, and M is the basis material density image.

[0145] Subtracting the low-energy image from the high-energy image yields equation (12).

[0146] I diff =M·A diff (12)

[0147] Among them, I diff Represents the image difference between high-energy image and low-energy image, A diff Represents the difference in decomposition matrices between high-energy and low-energy images.

[0148] Then, forward projection is performed on both sides of equation (12) at the same time. The symbol of forward projection is denoted by R. Then, equation (13) can be obtained after forward projection.

[0149] R.I. diff =R·M·A diff (13)

[0150] Among them, R.I diff Represents the energy data difference between high-energy image and low-energy image; R·M represents the material density data; A diff In some embodiments, A diff The least squares fitting method can be used to solve equation (13).

[0151] Based on the principles described above, in some embodiments, the processing device can obtain the material decomposition matrix difference based on the above equation (13). Specifically, the sample topology data difference is R·I diff , which represents the energy difference between the sample high-energy topological data and the sample low-energy topological data. The material density data can be obtained by forward projecting the sample basis material density image, that is, R·M. Finally, the material decomposition matrix difference can be obtained by least squares fitting calculation.

[0152] Step 608 : determining image differences based on the material decomposition matrix differences and the base material density image.

[0153] Image difference refers to the difference between the sample high-energy image and the sample low-energy image. After obtaining the material decomposition matrix difference through the sample topology data difference and the label base material density image, the image difference can be obtained by the equation (12) described above. In equation (12), M is the base material density image, A diff To obtain the material decomposition matrix difference, substitute the two into equation (12) to obtain the image difference I diff .

[0154] Step 610: Determine the topological high-energy image based on the sample low-energy image and the image difference.

[0155] In some embodiments, the processing device can determine the topological high-energy image based on the sample low-energy image and the image difference, and equations (10) to (12). diff =I High -I Low Already obtained, M.A diff I have also obtained Low is the sample low-energy image, then the topological high-energy image can be obtained by adding the image difference to the sample low-energy image.

[0156] It should be noted that the above descriptions of the various processes are intended for illustration and purpose only and do not limit the scope of this specification. Those skilled in the art may, under the guidance of this specification, make various modifications and alterations to the processes. However, such modifications and alterations remain within the scope of this specification. For example, changes to the process steps in this specification may include the addition of preprocessing and storage steps.

[0157] Figure 7 is an exemplary module diagram of a training system for an image processing model according to some embodiments of this specification. Figure 7 As shown, the system 700 may include a first acquisition module 710 , a first determination module 720 , and a parameter adjustment module 730 .

[0158] The first acquisition module 710 can be used to acquire multiple training samples.

[0159] In some embodiments, the first acquisition module 710 may obtain the multiple training sample labels by reading from a database, a storage device, or an imaging device.

[0160] The first determination module 720 may be configured to input the training sample into an image processing model to determine a base material density image corresponding to the sample low-energy image.

[0161] The parameter adjustment module 730 can be used to adjust the parameters of the image processing model based on the label base material density image, the sample low-energy image, the sample high-energy image, one or more of the sample topology data and the base material density image, with the minimization of the target loss function as the training goal, to obtain a trained image processing model.

[0162] The sample high-energy image corresponds to the sample low-energy image.

[0163] In some embodiments, the loss function corresponding to each training sample includes any one or a combination of a base material density image loss function, a low-energy image loss function, a first high-energy image loss function, and a second high-energy image loss function. The base material density image loss function is determined at least based on the labeled base material density image, the low-energy image loss function is determined at least based on the sample low-energy image, the first high-energy image loss function is determined at least based on the sample high-energy image, and the second high-energy image loss function is determined at least based on the sample high-energy image and sample topology data. The sample topology data includes sample low-energy topology data and sample high-energy topology data, and the sample low-energy topology data corresponds to the sample high-energy topology data.

[0164] Figure 8FIG. 1 is an exemplary module diagram of a system for generating a density image of a base material according to some embodiments of this specification. Figure 8 As shown, the system 800 may include a second acquisition module 810 and a second determination module 820 .

[0165] The second acquisition module 810 can be used to acquire an image to be processed of the target object.

[0166] The second determination module 820 can be used to input the image to be processed into the image processing model trained by the image processing model training method shown in the embodiments of this specification, and determine the base material density image of the image to be processed.

[0167] Figure 9 FIG is an exemplary module diagram of a system for generating high-energy images according to some embodiments of this specification. Figure 9 As shown, the system 900 may include a third acquisition module 910 , a third determination module 920 , a fourth determination module 930 , and a fifth determination module 940 .

[0168] The third acquisition module 910 may be configured to acquire low-energy images and topological data of the target object.

[0169] The topological data includes low-energy topological data and high-energy topological data, and the low-energy topological data corresponds to the high-energy topological data.

[0170] The third determination module 920 can be used to input the low-energy image into the image processing model trained by the image processing model training method described in the embodiments of this specification to determine the base material density image of the low-energy image.

[0171] The fourth determining module 930 may be configured to determine a topological data difference based on the low-energy topological data and the high-energy topological data.

[0172] The fifth determination module 940 may be configured to determine a high-energy image corresponding to the low-energy image based on the low-energy image, the base material density image, and the topological data difference.

[0173] For detailed description of each module of the above system, please refer to the flowchart section of this manual, for example, Figures 2 to 6 Related instructions.

[0174] It should be understood that Figure 7-Figure 9The system and its modules shown can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated hardware. Those skilled in the art will understand that the above-mentioned methods and systems can be implemented using computer-executable instructions and / or contained in processor control code, for example, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the system and its modules of this specification. Not only can the hardware circuits such as ultra-large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc. be implemented, they can also be implemented using software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software (for example, firmware).

[0175] It should be noted that the above description of the training system and modules of the image processing model is for convenience of description only and does not limit this specification to the scope of the embodiments cited. It is understandable that for those skilled in the art, after understanding the principle of the system, it is possible to arbitrarily combine the modules or form a subsystem to connect with other modules without deviating from this principle. For example, in some embodiments, the first acquisition module 710 and the first determination module 720 can be different modules in a system, or a module can realize the functions of two or more modules mentioned above. For example, each module can share a storage module, or each module can have its own storage module. Such variations are all within the scope of protection of this specification.

[0176] The beneficial effects that may be brought about by the embodiments of this specification include but are not limited to: (1) combining deep learning technology, using low-energy images to train the image processing model, and the trained image processing model can use single-energy CT images to realize the material decomposition function; (2) using multiple methods to construct loss functions for constraints during model training, the robustness of the trained image processing model can be improved; (3) the hardware implementation of the embodiments of this specification is simple, because the single-energy CT image obtained by scanning the target object with low-dose rays can obtain the base material density image, which reduces the radiation dose to the target object, and the obtained base material density image can have a higher signal-to-noise ratio; (4) before the low-energy image is input into the image processing model, it is not necessary to perform denoising and other processing on the low-energy image, which simplifies the process of obtaining the base material density image; (5) the output result of the image processing model is the base material density image. Compared with using a deep learning model to obtain an estimated high-energy image through a low-energy image, and then obtaining a base material density image based on the low-energy image and the estimated high-energy image, it can avoid or reduce the serious image signal-to-noise ratio degradation caused by using matrix inversion for the low-energy image and the high-energy image, and save the time required for iterative material decomposition.

[0177] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or any other possible beneficial effects.

[0178] While the basic concepts have been described above, it will be apparent to those skilled in the art that the detailed disclosure is merely illustrative and does not limit this specification. Although not explicitly stated herein, various modifications, improvements, and revisions to this specification may be made by those skilled in the art. Such modifications, improvements, and revisions are suggested in this specification and remain within the spirit and scope of the exemplary embodiments of this specification.

[0179] This specification also uses specific terms to describe the embodiments of this specification. For example, "one embodiment," "an embodiment," and / or "some embodiments" refer to a feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "one embodiment," "an embodiment," or "an alternative embodiment" two or more times in different locations in this specification do not necessarily refer to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of this specification may be appropriately combined.

[0180] In addition, it will be understood by those skilled in the art that various aspects of this specification may 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 this specification may be performed 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 "data blocks", "modules", "engines", "units", "components" or "systems". In addition, various aspects of this specification may be represented as a computer product located in one or more computer-readable media, which includes computer-readable program code.

[0181] A computer storage medium may include a propagated data signal embodying the computer program code, for example, in baseband or as part of a carrier wave. The propagated signal may be in a variety of forms, including electromagnetic, optical, or any suitable combination thereof. A computer storage medium may be any computer-readable medium other than a computer-readable storage medium that can be connected to an instruction execution system, apparatus, or device to communicate, propagate, or transfer the program for use. The program code on the computer storage medium may be transmitted via any suitable medium, including radio, cable, fiber optic cable, RF, or similar media, or any combination of these.

[0182] The computer program code required for the operation of the various parts of this specification 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, 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 stand-alone 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 via any network, 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).

[0183] In addition, unless expressly stated in the claims, the order of the processing elements and sequences, the use of alphanumeric characters, or the use of other names described in this specification are not intended to limit the order of the processes and methods of this specification. Although the above disclosure discusses some of the invention embodiments currently considered useful through various examples, it should be understood that such details are for illustrative purposes only, and the appended 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 spirit and scope of the embodiments of this specification. 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.

[0184] Similarly, it should be noted that, in order to simplify the presentation of this specification and thus facilitate understanding of one or more embodiments of the invention, the foregoing descriptions of the embodiments of this specification sometimes combine multiple features into a single embodiment, figure, or description thereof. However, this disclosure method does not imply that the subject matter of this specification requires more features than those recited in the claims. In fact, an embodiment may have fewer features than all of the features of a single disclosed embodiment.

[0185] In some embodiments, numbers are used to describe the quantity of components and attributes. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise stated, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the description and claims are approximate values, which may change according to the required characteristics of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of this specification are approximate values, in specific embodiments, the settings of such numerical values ​​are as accurate as possible within the feasible range.

[0186] Each patent, patent application, patent application publication, and other materials, such as articles, books, specifications, publications, and documents, cited in this specification is hereby incorporated by reference in its entirety. This includes application history documents that are inconsistent with or conflict with the content of this specification, as well as documents (currently or subsequently attached to this specification) that limit the broadest scope of the claims of this specification. It should be noted that if the descriptions, definitions, and / or terminology used in the accompanying materials are inconsistent or conflicting with the content of this specification, the descriptions, definitions, and / or terminology used in this specification will control.

[0187] Finally, it should be understood that the embodiments described in this specification are intended only to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly described and illustrated in this specification.

Claims

1. A training method for an image processing model, characterized in that: The method comprises: Acquire a plurality of training samples, wherein each of the plurality of training samples comprises a sample low-energy image; Inputting the training sample into an image processing model to determine a base material density image corresponding to the sample low-energy image; Based on the label base material density image, the sample low-energy image, the sample high-energy image, one or more of the sample topology data and the base material density image, the parameters of the image processing model are adjusted with the minimization of the target loss function as the training goal to obtain a trained image processing model; wherein the sample high-energy image corresponds to the sample low-energy image; the sample low-energy image is obtained by scanning the target object at a first radiation dose, and the sample high-energy image is obtained by scanning the target object at a second radiation dose, and the first radiation dose is lower than the second radiation dose.

2. The method according to claim 1, characterized in that The objective loss function includes: Any one or combination of a base material density image loss function, a low-energy image loss function, a first high-energy image loss function, and a second high-energy image loss function; Wherein, the base material density image loss function is determined at least based on the label base material density image, the low-energy image loss function is determined at least based on the sample low-energy image, the first high-energy image loss function is determined at least based on the sample high-energy image, and the second high-energy image loss function is determined at least based on the sample high-energy image and sample topology data; The sample topology data includes sample low-energy topology data and sample high-energy topology data, and the sample low-energy topology data corresponds to the sample high-energy topology data.

3. The method according to claim 2, characterized in that Determining a second high-energy image loss function based on the sample high-energy image and the sample topology data includes: determining a topological high-energy image based on the base material density image, the sample low-energy image, and the sample topological data; The second high-energy image loss function is determined based on the sample high-energy image and the topological high-energy image.

4. The method according to claim 3, characterized in that The determining of a topological high-energy image based on the base material density image, the sample low-energy image, and the sample topological data includes: determining sample material density data based on the base material density image; Determining a sample topological data difference based on the sample low-energy topological data and the sample high-energy topological data; Determining a material decomposition matrix difference based on a difference between the sample material density data and the sample topology data; determining an image difference based on the material decomposition matrix difference and the base material density image; The topological high-energy image is determined based on the sample low-energy image and the image difference.

5. A training system for an image processing model, characterized in that: The system comprises: A first acquisition module is configured to acquire a plurality of training samples, wherein each of the plurality of training samples comprises a sample low-energy image; A first determination module is configured to input the training sample into an image processing model to determine a base material density image corresponding to the sample low-energy image; A parameter adjustment module is used to adjust the parameters of the image processing model based on the label base material density image, the sample low-energy image, the sample high-energy image, one or more of the sample topology data and the base material density image, with minimizing the target loss function as the training goal, to obtain a trained image processing model; wherein the sample high-energy image corresponds to the sample low-energy image; the sample low-energy image is obtained by scanning the target object at a first radiation dose, and the sample high-energy image is obtained by scanning the target object at a second radiation dose, and the first radiation dose is lower than the second radiation dose.

6. A method for generating a density image of a base material, characterized in that The method comprises: Obtaining an image of the target object to be processed; The image to be processed is input into an image processing model trained by the method according to any one of claims 1 to 4 to determine a base material density image of the image to be processed.

7. A system for generating a density image of a base material, characterized in that The system comprises: A second acquisition module is used to acquire an image to be processed of the target object; The second determination module is used to input the image to be processed into the image processing model trained by the method according to any one of claims 1 to 4, and determine the base material density image of the image to be processed.

8. A method for generating a high-energy image, characterized in that: The method comprises: Acquire a low-energy image and topological data of a target object; the topological data includes low-energy topological data and high-energy topological data, and the low-energy topological data corresponds to the high-energy topological data; Inputting the low-energy image into an image processing model trained by the method according to any one of claims 1 to 4 to determine a base material density image of the low-energy image; determining a topological data difference based on the low-energy topological data and the high-energy topological data; A high-energy image corresponding to the low-energy image is determined based on the low-energy image, the base material density image, and the topological data difference.

9. A system for generating high-energy images, characterized in that: The system comprises: A third acquisition module is configured to acquire a low-energy image and topological data of the target object; the topological data includes low-energy topological data and high-energy topological data, and the low-energy topological data corresponds to the high-energy topological data; a third determining module, configured to input the low-energy image into an image processing model trained by the method according to any one of claims 1 to 4, and determine a base material density image of the low-energy image; a fourth determining module, configured to determine a topological data difference based on the low-energy topological data and the high-energy topological data; A fifth determining module is configured to determine a high-energy image corresponding to the low-energy image based on the low-energy image, the base material density image, and the topological data difference.

10. A training device for an image processing model, comprising at least one storage medium and at least one processor, wherein the at least one storage medium is used to store computer instructions; and the at least one processor is used to execute the computer instructions to implement the method according to any one of claims 1 to 4.

Images