Metal artifact correction method and system

Abstract

The embodiment of the invention discloses a metal artifact correction method. The correction method comprises the steps that a first image obtained by a first imaging device and a second image obtained by a second imaging device are obtained, and the second imaging device is a CT device; registering the first image and the second image; determining a template image based on the registered first image and second image; determining a metal artifact image based on the template image; and correcting the second image based on the metal artifact image to obtain a corrected image.

Description

Metal artifact correction method and system Technical Field

[0001] The present invention relates to the field of image processing, and in particular to a method and system for correcting metal artifacts in digital images. Background Art

[0002] With the development of radiotherapy technology, radiotherapy has entered the stage of precision radiotherapy, in which target marking, organ delineation and dose calculation through CT images are very critical steps in the radiotherapy process. However, when there are metal implants in the human body (for example, joint prostheses, tooth fillings, spinal fixation bolts, etc.), the metal has a very high absorption rate of X-rays, which will cause the detector to not receive enough X-ray signals, and ultimately form light and dark strip artifacts in the reconstructed CT image. The reconstructed image with metal artifacts is of poor quality, which will affect the application of subsequent images, especially in the case of multiple metals / complex metals. The artifacts caused by metal are very serious, making it difficult to accurately delineate the target area and organs, and the inaccuracy of the CT value in the image will also cause a large deviation in the dose during dose calculation.

[0003] Therefore, it is necessary to propose a metal artifact correction method for repairing CT images that are severely contaminated by metal artifacts.

[0004] Summary of the Invention

[0005] One embodiment of this specification provides a metal artifact correction method. The method includes: acquiring a first image obtained by a first imaging device and a second image obtained by a second imaging device, wherein the second imaging device is a CT device; registering the first image and the second image; determining a template image based on the registered first and second images; determining a metal artifact image based on the template image; and correcting the second image based on the metal artifact image to obtain a corrected image.

[0006] One embodiment of this specification provides a metal artifact correction system. The system includes: an acquisition module for acquiring a first image obtained by a first imaging device and a second image obtained by a second imaging device, wherein the second imaging device is a CT device; a registration module for registering the first image and the second image; a template image determination module for determining a template image based on the registered first and second images; a metal artifact determination module for determining a metal artifact image based on the template image; and a correction module for correcting the second image based on the metal artifact image to obtain a corrected image.

[0007] One embodiment of this specification provides a metal artifact correction system. The system includes: at least one storage device for storing computer instructions; and at least one processor configured to communicate with the at least one storage device, wherein when executing the set of instructions, the at least one processor is configured to perform the following operations. The operations include: acquiring a first image obtained by a first imaging device and a second image obtained by a second imaging device, wherein the second imaging device is a CT device; registering the first image and the second image; determining a template image based on the registered first and second images; determining a metal artifact image based on the template image; and correcting the second image based on the metal artifact image to obtain a corrected image.

[0008] One embodiment of this specification provides a metal artifact correction device. The device includes at least one processor and at least one memory, wherein the at least one memory is configured to store computer instructions, and the at least one processor is configured to execute at least some of the computer instructions to implement the metal artifact correction method described in any embodiment of this specification.

[0009] One embodiment of this specification provides a computer-readable storage medium, wherein the storage medium stores computer instructions. When the computer instructions are executed by a processor, the metal artifact correction method as described in any embodiment of this specification is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] 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:

[0011] FIG1 is a schematic diagram of an application scenario of a metal artifact correction system according to some embodiments of this specification;

[0012] FIG2 is a module diagram of a metal artifact correction system according to some embodiments of this specification;

[0013] FIG3 is an exemplary flow chart of a metal artifact correction method according to some embodiments of this specification;

[0014] 4A and 4B are schematic diagrams of a first image and a second image according to some embodiments of this specification;

[0015] FIG5 is an exemplary flowchart of determining a template image according to some embodiments of this specification;

[0016] FIG6 is an exemplary flowchart of determining a template image according to some embodiments of this specification;

[0017] FIG7 is an exemplary flowchart of determining a metal artifact image according to some embodiments of this specification;

[0018] FIG8 is an exemplary flowchart of metal artifact correction according to some embodiments of the present specification; and

[0019] FIG. 9 is a diagram comparing the effects of metal artifact correction methods according to some embodiments of this specification. DETAILED DESCRIPTION

[0020] 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.

[0021] 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, if other terms can achieve the same purpose, the terms may be replaced by other expressions.

[0022] 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.

[0023] 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.

[0024] As previously mentioned, when metal implants are present in the human body, due to the metal's high X-ray absorption rate, when X-rays pass through the metal, they are almost completely absorbed. This results in insufficient X-ray signal being received by the detector, and the back-projection value is abnormally large. When the tomographic ray attenuation coefficient exhibits discontinuous jumps, the first-order derivative of the projection data exhibits weak continuity at certain intervals. After filtering, this weak continuity is further amplified, resulting in alternating light and dark stripe artifacts in the reconstructed image.

[0025] The embodiments of this specification provide a metal artifact correction method and system. The metal artifact correction method disclosed in the embodiments of this specification can be applied to CT images containing metal artifacts. For example, for a kilovolt CT image, the metal artifact portion of the kilovolt CT image can be determined by combining it with an image with less metal artifacts (e.g., a megavolt CT image), thereby repairing a kilovolt CT image that is severely contaminated by metal artifacts.

[0026] Specifically, the method can be implemented by acquiring a first image obtained by a first imaging device and a second image obtained by a second imaging device, wherein the second imaging device is a CT device; registering the first image and the second image; determining a template image based on the registered first and second images; determining a metal artifact image based on the template image; and correcting the second image based on the metal artifact image to obtain a corrected image. The metal artifact correction method described in the embodiments of this specification allows the first image and the second image to be registered, and then the template image is obtained by using the geometric information of different tissue regions in the first image and the CT value information in the second image, and then the metal artifact image is generated, thereby ultimately achieving image restoration. This method can achieve good restoration effects even in the case of multiple metals / large metals, and can ensure the accuracy of the image CT value while restoring the image geometry, thus meeting the requirements of clinical target area delineation and dose calculation.

[0027] FIG1 is a schematic diagram of an application scenario of a metal artifact correction system according to some embodiments of this specification.

[0028] As shown in FIG. 1 , the metal artifact correction system 100 may include an imaging device 110 , a network 120 , a terminal 130 , a processing device 140 , and a storage device 150 .

[0029] The imaging device 110 can use different media to reproduce the structure of the target object as a specific medical image. In some embodiments, the target object may include a biological object and / or a non-biological object. For example, the target object may include a specific part of the human body, such as the neck, chest, abdomen, etc., or a combination thereof. At least one metal implant is present inside the target object. In some embodiments, the imaging device 110 can be a medical device that uses a variety of imaging technologies, such as a digital subtraction angiography (DSA) device, a computed radiography (CR) system, a digital radiography (DR) system, a computed tomography (CT) device, an ultrasound imaging device, a fluoroscopic imaging device, a magnetic resonance imaging (MRI) device, a positron emission tomography (PET) device, etc., or any combination thereof. In some embodiments, the imaging device 110 can include a first imaging device and a second imaging device, which can be used to respectively acquire a first image and a second image of the target object, wherein the first image is an image with relatively weak metal artifacts, and the second image is a CT image. In some embodiments, the first imaging device is a first energy level CT device, and the second imaging device is a second energy level CT device, wherein the first energy level is greater than the second energy level. In some embodiments, the first energy level CT device is a megavolt CT device, and the second energy level CT device is a kilovolt CT device. In some embodiments, the first imaging device and the second imaging device can be two independent imaging devices, or two imaging devices on the same imaging device, each performing one of the functions. The imaging device 110 provided above is for illustrative purposes only and is not intended to limit its scope.

[0030] In some embodiments, the imaging device 110 may be part of a radiotherapy device, which may include the imaging device 110, a radiotherapy device, a main frame, and a support plate (not numbered in the figure). A target object (e.g., a patient) lying on the support plate may be imaged by the imaging device, and then the patient may be moved directly and linearly to the radiotherapy device for radiotherapy. This reduces positioning errors caused by patient movement and improves the efficiency of the radiation treatment process.

[0031] In some embodiments, the imaging device 110 may include a radiation emitting device and a detection device. The radiation emitting device may emit radioactive rays that pass through a target object and are received by the detection device. The radioactive rays may include one or a combination of corpuscular rays, photon rays, etc. The corpuscular rays may include one or a combination of neutrons, protons, electrons, muon media, heavy ions, etc. The photon rays may include one or a combination of X-rays, gamma rays, alpha rays, beta rays, ultraviolet rays, lasers, etc. By way of example only, the photon rays may be X-rays, and the radiation emitting device may be an X-ray tube or an accelerator tube.

[0032] In some embodiments, the imaging device 110 can acquire a first image acquired by a first imaging device and a second image acquired by a second imaging device, and send the first image and the second image to the processing device 140. In some embodiments, the first imaging device is a megavoltage CT device, and the second imaging device is a kilovoltage CT device. For example, the first image is a megavoltage cone-beam CT image, and the second image is a kilovoltage diagnostic CT image. In some embodiments, the image acquired by the imaging device 110 can be stored in the storage device 140. In some embodiments, the imaging device 110 can receive an imaging instruction sent from a terminal (not shown) or the processing device 140 via the network 120, and can send the imaging result to the processing device 140 or the storage device 140. In some embodiments, one or more components of the metal artifact correction system 100 (e.g., the processing device 140, the storage device 140) can be included in the imaging device 110.

[0033] The network 120 may include any suitable network that facilitates the exchange of information and / or data between the metal artifact correction system 100. In some embodiments, one or more other components of the metal artifact correction 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 each other via the network 120. For example, the processing device 140 may obtain the first image and / or the second image from the imaging device 110 via the network 120. For another example, the processing device 140 may obtain a user instruction from the terminal 130 via the network 120. The network 120 may be and / or include 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., Ethernet), a wireless network (e.g., an 802.11 network, a Wi-Fi network, etc.), a cellular network (e.g., an LTE network), a frame relay network, a virtual private network ("VPN"), a satellite network, a telephone network, a router, a hub, a switch, a server computer, and / or a combination of one or more thereof. For example, the network 120 may include a cable network, 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 a wired and / or wireless network access point, such as a base station and / or a network switching point, through which one or more components of the system 100 may access the network 120 to exchange data and / or information.

[0034] In some embodiments, a user can operate the metal artifact correction system 100 through the terminal 130. The terminal 130 may include a combination of one or more of a mobile device 131, a tablet computer 132, a laptop computer 133, and the like. In some embodiments, the corrected image may be presented to the user through the terminal 130. In some embodiments, the mobile device 131 may include a combination of one or more of a smart home device, a wearable device, a mobile device, a virtual reality device, an augmented reality device, and the like. In some embodiments, the smart home device may include a combination of one or more of a smart lighting device, a smart appliance control device, a smart monitoring device, a smart TV, a smart camera, an intercom, and the like. In some embodiments, the wearable device may include a combination of one or more of a bracelet, shoes and socks, glasses, a helmet, a watch, clothing, a backpack, smart accessories, and the like. In some embodiments, the mobile device may include a combination of one or more of a mobile phone, a personal digital assistant (PDA), a gaming device, a navigation device, a point of sale (POS) device, a laptop computer, a tablet computer, a desktop computer, and the like. In some embodiments, the virtual reality device and / or augmented reality device may include a combination of one or more of a virtual reality helmet, virtual reality glasses, virtual reality goggles, an augmented reality helmet, augmented reality glasses, augmented reality goggles, and the like. For example, the virtual reality device and / or augmented reality device may include Google Glass. TM 、Oculus Rift TM 、Hololens TM 、Gear VR TM In some embodiments, the terminal 130 may be a part of the processing device 140 . In some embodiments, the terminal 130 may be a part of the imaging device 110 .

[0035] 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 obtain a first image and a second image from the imaging device 110 and perform registration processing on the first image and the second image to generate a template image. For another example, the processing device 140 can determine a metal artifact image based on the template image and correct the second image based on the metal artifact image. In some embodiments, the processing device 140 can be a 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 stored in the imaging device 110, the terminal 130, and / or the storage device 150 via the network 120. For example, the processing device 140 can directly connect to the imaging device 110, the terminal 130, and / or the storage device 150 to access the information and / or data stored therein. In some embodiments, the processing device 140 can be executed on a cloud platform. For example, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an interconnected cloud, a multi-cloud, or a combination thereof. In some embodiments, the processing device 140 may be executed by a computing device having one or more components. In some embodiments, the processing device 140 may be part of the imaging device 110 or the terminal 130.

[0036] The storage device 150 can store data, instructions, and / or other information. In some embodiments, the storage device 150 can store data obtained from the terminal 130 and / or the processing device 140. In some embodiments, the storage device 150 can store data and / or instructions executed or used by the processing device 140 to perform the exemplary methods described herein. In some embodiments, the storage device 150 can include a combination of one or more of mass storage, removable storage, volatile read-write memory, read-only memory (ROM), and the like. Exemplary mass storage can include magnetic disks, optical disks, solid-state drives, and the like. Exemplary removable storage can include flash drives, floppy disks, optical disks, memory cards, ziplock disks, magnetic tapes, and the like. Exemplary volatile read-write memory can include random access memory (RAM). Exemplary RAM can 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), and zero capacitance random access memory (Z-RAM). Exemplary read-only memories (ROMs) may include mask read-only memories (MROMs), programmable read-only memories (PROMs), erasable programmable read-only memories (EPROMs), electrically erasable programmable read-only memories (EEPROMs), compact disc read-only memories (CD-ROMs), and digital versatile discs. In some embodiments, the storage device 150 may be executed on a cloud platform. For example, the cloud platform may include a combination of one or more of a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an interconnected cloud, or multiple clouds.

[0037] In some embodiments, the storage device 150 can be connected to the network 120 to communicate with one or more other components of the system 100 (e.g., the processing device 140, the terminal 130, etc.). One or more components of the metal artifact correction system 100 can access data or instructions stored in the storage device 150 through the network 120. In some embodiments, the storage device 150 can be directly connected to or communicate with one or more other components of the system 100 (e.g., the processing device 140, the terminal 130, etc.). In some embodiments, the storage device 150 can be part of the processing device 140.

[0038] The above description of the metal artifact correction system 100 is merely exemplary. In some embodiments, the metal artifact correction system 100 may include one or more other components, or may not include one or more of the components described above. Alternatively, two or more components may be combined into one component.

[0039] FIG2 is a module diagram of a metal artifact correction system 200 according to some embodiments of this specification.

[0040] 2 , the metal artifact correction system 200 may include an acquisition module 210, a registration module 220, a template image determination module 230, a metal artifact determination module 240, and a correction module 250. In some embodiments, the acquisition module 210, the registration module 220, the template image determination module 230, the metal artifact determination module 240, and the correction module 250 may be implemented by the processing device 140.

[0041] Acquisition module 210 can be used to acquire a first image obtained by a first imaging device and a second image obtained by a second imaging device, wherein the second imaging device is a CT device. In some embodiments, the first imaging device can be a first energy level CT device, and the second imaging device can be a second energy level CT device, where the first energy level is greater than the second energy level. In some embodiments, the first energy level device can be a megavolt CT device, and the second energy level CT device can be a kilovolt CT device. In some embodiments, the first image can be a cone-beam CT image, and the second image can be a diagnostic-grade CT image. For more information on acquiring the first and second images, please refer to the detailed description of step 310 and are not further elaborated here.

[0042] The registration module 220 may be used to register the first image and the second image. For more details on registering the first image and the second image, please refer to the detailed description of step 320, which will not be repeated here.

[0043] The template image determination module 230 can be configured to determine a template image based on the registered first and second images. In some embodiments, the template image determination module 230 can determine regions of varying density based on the first image and fill these regions of varying density based on the CT values ​​of regions corresponding to these regions in the second image to obtain a template image. For more information on determining the template image, please refer to the detailed description of step 330 and FIG. 5 and are not further elaborated here.

[0044] The metal artifact determination module 240 can be used to determine a metal artifact image based on a template image. In some embodiments, the metal artifact determination module 240 can determine a metal area based on the second image and perform a forward projection on the metal area to generate a metal projection map, wherein the non-zero area in the metal projection map is the metal projection area. In some embodiments, the metal artifact determination module 240 can perform a forward projection on the second image to generate a second image projection map. In some embodiments, the metal artifact determination module 240 can perform a forward projection on the template image to generate a template image projection map. In some embodiments, the metal artifact determination module 240 can determine the difference in projection map between the second image projection map and the template image projection map in the metal projection area, and determine the metal artifact image based on the projection map difference. For more information on determining the metal artifact image, please refer to the detailed description of step 340 and Figure 7, which will not be repeated here.

[0045] The correction module 250 can be configured to correct the second image based on the metal artifact image to obtain a corrected image. In some embodiments, the correction module 250 can filter the metal artifact image to obtain a processed metal artifact image, and correct the second image based on the processed metal artifact image to obtain a corrected image. In some embodiments, the correction module 250 can subtract the metal artifact image from the second image to obtain the corrected image. In some embodiments, the correction module 250 can update the second image based on the corrected image for one or more iterations. For more information on obtaining the corrected image, please refer to the detailed description of step 350 and will not be repeated here.

[0046] It should be understood that the system and its modules shown in Figure 2 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.

[0047] It should be noted that the above description of the system and its modules is for convenience of description only and does not limit this specification to the scope of the embodiments. 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 acquisition module 210, the registration module 220, the template image determination module 230, the metal artifact determination module 240 and the correction module 250 disclosed in Figure 2 can be different modules in a system, or a module can implement the functions of two or more of the above modules. For example, the modules 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.

[0048] FIG3 is an exemplary flowchart of a metal artifact correction method according to some embodiments of this specification.

[0049] Process 300 can be executed by a processing device (e.g., processing device 140). For example, process 300 can be implemented as an instruction set (e.g., an application program) stored in a memory internal to or external to metal artifact correction system 100. The processing device can execute the instruction set and, when executing the instructions, can be configured to perform process 300. The operational diagram of process 300 presented below is illustrative. In some embodiments, the process can be completed using one or more additional operations not described above and / or omitting one or more operations discussed below. In addition, the order of the operations of process 300 shown in FIG3 and described below is not intended to be limiting.

[0050] As described elsewhere in this specification, when metal implants are present in the human body, the metal's high X-ray absorption rate prevents the detector from receiving sufficient X-ray signals, ultimately resulting in the formation of alternating light and dark stripe artifacts in the reconstructed CT image. Reconstructed images with metal artifacts are of poor quality, impacting subsequent image applications. This is particularly true for multi-metal / complex metal cases, where artifacts caused by metal are extremely severe, making it difficult to accurately delineate target areas and organs. Furthermore, inaccuracies in CT values ​​in the image can lead to significant dose deviations during dose calculation. Therefore, a metal artifact correction method is needed that achieves good restoration results even in multi-metal / large metal cases, ensuring the accuracy of CT values ​​while restoring image geometry, and meeting the requirements of clinical target delineation and dose calculation.

[0051] Step 310: Acquire a first image obtained by a first imaging device and a second image obtained by a second imaging device. The second imaging device is a CT device. In some embodiments, step 310 may be performed by the processing device 140 or the acquisition module 210.

[0052] The first image and the second image are anatomical images acquired from the same target object, reflecting the internal structure of the target object. In some embodiments, the first image and the second image may include anatomical images of the entire target object, or may only include anatomical images of a portion of the target object, such as the neck, chest, or abdomen. The first image is an image with minimal metal artifacts and can be used to assist in metal artifact correction for the second image, while the second image is a CT image for which metal artifact correction is required. In some embodiments, the first imaging device may be a CT device, and the first image and the second image may be CT images acquired using CT devices with different energy levels. Specifically, the first imaging device may be a first energy level CT device, and the second imaging device may be a second energy level CT device, with the first energy level being greater than the second energy level. In some embodiments, the first image and the second image may be pre-captured and stored. If, during the acquisition of the first and second images, the position of the target object is identical or substantially identical in both images, the position and state of the metal implant and the target object in the first and second images may be substantially identical, facilitating subsequent registration of the first and second images to determine the metal artifact image.

[0053] In some embodiments, the first energy level CT device may be a megavolt CT device, and the second energy level CT device may be a kilovolt CT device. The megavolt level may include 1 to 999 MV, and the kilovolt level may include 1 to 999 KV. In some embodiments, the kilovolt CT device may include a 70 to 140 KV CT device, and the megavolt CT device may include a 1 to 6 MV CT device. In some embodiments, the first energy level CT device and the second energy level CT device may both be a cone beam CT device or a fan beam CT device. In some embodiments, the first image may be a cone beam CT image, and the second image may be a diagnostic CT image. In an embodiment of the present specification, the first image may be a megavolt cone beam CT image (MV-CBCT image), and the second image may be a kilovolt CT image (kV-CT image). In some embodiments, the first energy level CT device and the second energy level CT device may be two independent CT devices, or two imaging devices on the same imaging device, each having one of the functions. In some embodiments, the first energy level CT device and the second energy level CT device can share a support plate. By moving the support plate, the target object can be positioned on the first energy level CT device and the second energy level CT device respectively. The first image and the second image can be acquired without taking the target object offline, thereby reducing position errors, improving the accuracy of registration, and further improving the artifact correction effect.

[0054] The first image and the second image are both CT images. To reconstruct the CT image, an imaging device (e.g., a first-level CT device and a second-level CT device) scans a target object using rays (e.g., X-rays) emitted by a ray emitting device. The rays that have passed through the target object are then received by a detection device and converted into digital signals that are processed to obtain a number of voxels. The X-ray attenuation coefficient or absorption coefficient of each voxel is arranged into a digital matrix, which is then converted into a pixel matrix to obtain the reconstructed CT image.

[0055] In some embodiments, an MV-CBCT image is an image reconstructed after a CBCT (Cone beam CT) scan using megavoltage rays generated by a linear accelerator, and a kV-CT image is an image reconstructed after a CT scan using kilovoltage rays generated by a ray tube.

[0056] During imaging, megavoltage radiation has greater penetrability than kilovoltage radiation, reducing metal artifacts in reconstructed images. However, images obtained using megavoltage radiation inherently have large CT value deviations and low density resolution, making them inadequate for direct replacement of kilovoltage CT images for subsequent applications. Therefore, this document considers combining megavoltage and kilovoltage CT images for metal artifact correction.

[0057] In some embodiments, processing device 140 may also use other high-energy images to correct the second image, as long as the energy difference with the second image is significant. For example, CT images of other energy levels may be used to correct the second image. For example, the first image may be a megavoltage CT image (greater than 1 MV), while the second image is still a kilovoltage CT image. In some embodiments, processing device 140 may also use MR images, PET images, ultrasound images, and other images with no or very weak metal artifacts to correct the second image. Specifically, the first imaging device may be an MR device, PET device, or ultrasound device.

[0058] In some embodiments, the acquisition module 210 may directly acquire the first image and / or the second image from the imaging device 110 , or may acquire the stored first image and / or the second image from the storage device 150 .

[0059] Step 320 : Register the first image and the second image. In some embodiments, step 320 may be performed by the processing device 140 or the registration module 220 .

[0060] To help determine the metal artifact image in the second image, the first image and the second image need to be registered using an image registration method to help generate a template image. Specifically, the image registration method can include a combination of one or more of a registration method based on grayscale information or a template, a registration method based on a transform domain, and a registration method based on a feature.

[0061] Because both the first image and the second image can be represented by a pixel matrix, in some embodiments, the processing device 140 can convert the first image into a pixel matrix of the same size as the pixel matrix of the second image based on the second image (e.g., the pixel matrix of the second image). As an example, the first image can be an MV-CBCT image and the second image can be a kV-CT image. If the pixel matrix of the second image is an N*P*Q matrix, the pixel matrix of the first image can also be converted into an N*P*Q matrix.

[0062] In some embodiments, the processing device 140 may register the first image and the second image using a rigid registration method, which may include a translation transformation and a rotation transformation. Specifically, during the rigid registration process, the processing device 140 may select corresponding pixels on the first image and the second image to determine relevant parameters of the rigid registration, such as the amount of translation and rotation of the images, thereby registering the first image and the second image.

[0063] In some embodiments, the processing device 140 may register the first image and the second image with the help of a deep learning network, such as a registration model. The deep learning network can significantly shorten the registration time and improve the registration accuracy with the help of the reasoning ability of the model. In some embodiments, the deep learning network may include but is not limited to a combination of one or more of a radial basis function neural network, a deep neural network, a convolutional neural network, a recursive neural network, a generative adversarial network, etc. In some embodiments, the processing device 140 may input the first image and the second image (for example, the pixel matrices of the first image and the second image) based on the trained registration model to output the registration result of the first image and the second image (for example, the overall translation and rotation of the image).

[0064] Step 330 : Determine a template image based on the registered first image and second image. In some embodiments, step 330 may be performed by the processing device 140 or the template image determination module 230 .

[0065] The template image is an intermediate processed image used to generate a metal artifact image based on the first and second images. In some embodiments, processing device 140 can determine different density regions based on the first image. These different density regions can include a high-density region M1, a medium-density region M2, and a low-density region M3. The high-density region M1 can correspond to bone tissue and implanted metal areas, the medium-density region M2 can correspond to soft tissue areas, and the low-density region M3 can correspond to internal cavities and external air areas.

[0066] In some embodiments, the processing device 140 may process the first image based on a local threshold segmentation method to determine high-density areas, medium-density areas, and low-density areas. In some embodiments, the processing device 140 may also process the first image based on a global threshold segmentation method to determine high-density areas, medium-density areas, and low-density areas. In some embodiments, the processing device 140 may process the first image based on a local threshold segmentation method to determine high-density areas and medium-density areas, and then determine the low-density area based on a composite image fused with the second image and the first image. Since some metal artifacts still exist in the first image (for example, an MV-CBCT image), resulting in an uneven distribution of CT values ​​in the image matrix, if a global threshold segmentation method is used, bone tissue and metal artifacts are easily classified into the same area, which is more prone to errors than the local threshold segmentation method. Therefore, in order to achieve a better optimization and correction effect, the local threshold segmentation method can be used as much as possible.

[0067] In some embodiments, the processing device 140 can fill in the different density regions based on the CT values ​​of the corresponding regions in the second image, thereby obtaining a template image. Specifically, for different density regions, filling can be performed directly based on the CT values ​​of the corresponding regions in the second image, or based on the average CT values ​​of each pixel. The CT value can be used to represent the attenuation coefficient or absorption coefficient of a material to radiation. The CT value of a material can reflect its density, i.e., a higher CT value corresponds to a higher density. The unit of the CT value is Hounsfield Unit (Hu). The CT value can be calculated by taking the difference between the attenuation coefficient of the material and the attenuation coefficient of water, comparing it with the attenuation coefficient of water, and then multiplying it by a certain scaling factor. Different tissues in the human body have different attenuation coefficients and, therefore, different CT values. By way of example only, tissues in the human body can be arranged from high to low CT values, such as bone tissue, soft tissue, and internal cavities. For example, the CT value of water can be around 0 Hu.

[0068] For more details on how to determine the template image, please refer to the detailed description of FIG5 , which will not be repeated here.

[0069] Step 340 : Determine a metal artifact image based on the template image. In some embodiments, step 340 may be performed by the processing device 140 or the metal artifact determination module 240 .

[0070] The metal artifact image is a portion of the reconstructed image that contains artifacts due to the influence of metal implants. As previously mentioned, the high absorption rate of radiation by metal implants results in alternating light and dark stripe artifacts in the reconstructed image. Megavoltage radiation, however, has greater penetrating power, which can reduce metal artifacts in the reconstructed image. Therefore, the intensities of metal artifacts in a first image obtained using a CT device with a first energy level (e.g., megavoltage) and a second image obtained using a CT device with a second energy level (e.g., kilovoltage) differ. This allows the use of a template image derived from the first and second images to determine the metal artifact image. The metal artifact image can be considered an image that isolates these stripe artifacts to a certain extent, thereby enabling correction of the second image.

[0071] In some embodiments, the processing device 140 can perform forward projection based on the second image and the template image, and determine the metal artifact image by the difference in the obtained projection images. Specifically, the processing device 140 can determine the metal area based on the second image, and perform forward projection on the metal area to generate a metal projection map, wherein the non-zero area in the metal projection map is the metal projection area. Further, the processing device 140 can perform forward projection on the second image to generate a second image projection map. The processing device 140 can further perform forward projection on the template image to generate a template image projection map. Then, the processing device 140 can determine the difference in projection map between the second image projection map and the template image projection map in the metal projection area. Finally, the processing device 140 can determine the metal artifact image based on the difference in projection map between the second image projection map and the template image projection map in the metal projection area.

[0072] For more details on how to determine the metal artifact image, please refer to the detailed description of FIG7 , which will not be repeated here.

[0073] Step 350 : Correct the second image based on the metal artifact image to obtain a corrected image. In some embodiments, step 350 may be performed by the processing device 140 or the correction module 250 .

[0074] In some embodiments, processing device 140 can subtract the metal artifact image from the second image to obtain a corrected image. To reduce the impact of the metal implant, the portion of the second image (e.g., a kV-CT image) affected by the metal implant and resulting in artifacts can be directly subtracted from the second image (e.g., a kV-CT image) for subsequent applications. The corrected image, thus representing the restored second image, significantly reduces the impact of the metal implant compared to the original second image, enabling better application in subsequent clinical applications, such as diagnostic processing, target delineation, and dose calculation.

[0075] Because the registration and threshold segmentation operations in steps 320 and 330 may not be sufficiently accurate, the resulting metal artifact image may not completely overlap with the second image. Directly subtracting the metal artifact image from the second image may produce new streak artifacts. Therefore, filtering the metal artifact image to remove high-frequency components before performing the correction process can be considered.

[0076] In some embodiments, before performing correction, the processing device 140 may first perform a certain degree of filtering on the metal artifact image to obtain a processed metal artifact image to reduce noise generated by artifact boundaries. Specifically, the processing device 140 may filter the metal artifact image using high-frequency filtering. Such high-frequency filtering methods may include, but are not limited to, one or more of mean filtering and Gaussian filtering. High-frequency filtering can be considered a blurring process. As an example, the processing device 140 may remove pixels in the image matrix of the metal artifact image whose CT values ​​exceed a pre-set filtering threshold. The filtering threshold can be set manually or adjusted based on the correction results. In some embodiments, the processing device 140 may filter the metal artifact image using machine learning to achieve high-frequency noise reduction. Specifically, the processing device 140 may input the metal artifact image into a trained machine learning model and directly output the processed metal artifact image. Compared to the unfiltered artifact image, the filtered image has a more blurred artifact boundary, less sharp edges, and a slower grayscale transition, effectively reducing errors that may have occurred in the previous steps.

[0077] In some embodiments, the processing device 140 may correct the second image based on the processed metal artifact image to obtain the corrected image. Specifically, the processing device 140 may directly subtract the processed metal artifact image from the second image to obtain the corrected image. This process can be expressed as the following formula (1):

[0078] I correct =I CT -HFF[Imetal )] (1)

[0079] Among them, I metal Represents the determined metal artifact image, HFF represents high frequency filtering, I CT Represents the original CT image, i.e., the second image, I correct represents the corrected image. As can be seen from formula (1), the corrected image can be obtained by performing high-frequency filtering on the metal artifact image and then subtracting the high-frequency filtered metal artifact image from the second image. For more information on determining the difference between the metal artifact image and the projection image, please refer to the detailed description of Figure 7 and will not be repeated here.

[0080] In some embodiments, the processing device 140 may also omit the step of filtering the metal artifact image and directly subtract the metal artifact image that has not been filtered from the second image to reduce the overall computational complexity.

[0081] During the complete calibration process described above, since the template image is determined using the CT values ​​in the second image, it also includes the influence of metal artifacts, which may cause inaccurate CT values ​​in the artifact areas of the calibrated image. Furthermore, the high-frequency filtering of the artifact image may also cause loss of artifact details. A single calibration may not completely remove the artifacts, and some artifacts may still remain in the calibrated image. Therefore, a template image can be regenerated based on the calibrated image and incorporated into the calibration process for iterative calibration.

[0082] In some embodiments, the processing device 140 may update the second image based on the corrected image to perform one or more iterations, i.e., may repeatedly perform steps 330 to 350. Specifically, the processing device 140 may update the template image based on the updated second image. For example, the template image may be filled based on the CT values ​​of corresponding regions in the updated second image to obtain an updated template image. Furthermore, the processing device 140 may update the metal artifact image based on the updated second image and the updated template image. For example, the updated template image and the updated second image may be forward-projected to generate an updated template image projection map and an updated second image projection map, and then an updated projection map difference between the updated template image projection map and the updated second image projection map may be determined. An updated metal artifact image may then be determined based on the updated projection map difference. Since the location of the metal implant is fixed, there is no need to repeatedly determine the metal area and the metal projection area. Finally, the processing device 140 may correct the updated second image based on the updated metal artifact image to obtain an updated corrected image. For example, after filtering the updated metal artifact image, the updated metal artifact image is subtracted from the updated second image to obtain an updated corrected image.

[0083] By performing one or more rounds of iterations, metal artifacts can be gradually removed during each iteration to obtain a better correction effect. In some embodiments, the number of iterations can be set manually, or it can be adjusted based on various parameter settings and correction results during the calculation process. For example, the number of iterations can be set to 5 times, and the corrected image obtained after repeating steps 330 to 350 for 5 times can be regarded as the final corrected image, and it is believed that the image has a good repair effect and can be subsequently used for clinical applications, such as clinical target area delineation and dose calculation. In some embodiments, the iterative correction process can be optimized through a machine learning process. For example, the number of iterations is adaptively set through a machine learning model.

[0084] The metal artifact correction method provided in the embodiments of this specification achieves image restoration by registering a first image at a first energy level with a second image at a second energy level. The method then utilizes the geometric information of different tissue regions in the first image and the CT value information in the second image to generate a template image, which then generates a metal artifact image. Furthermore, it can be further iterated based on the corrected image to update the second image, gradually removing artifacts and improving the image correction effect. This method achieves excellent restoration results even in the presence of multiple or large metals, maintaining the accuracy of the image CT values ​​while restoring the image geometry, meeting the requirements of clinical target delineation and dose calculation.

[0085] It should be noted that the above description of process 300 is for illustration and purpose only and does 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 process 300. However, such modifications and alterations remain within the scope of this specification. For example, the process of steps 320 to 350 may be repeated.

[0086] 4A and 4B are schematic diagrams of a first image and a second image according to some embodiments of the present specification.

[0087] Figure 4A is a schematic diagram of a first image, which is an MV-CBCT image. Figure 4B is a schematic diagram of a second image, which is a kV-CT image. A comparison of Figures 4A and 4B shows that, due to the greater penetrability of megavoltage radiation, metal artifacts in MV-CBCT images are much weaker than those in kV-CT images. Therefore, MV-CBCT images can be used to process kV-CT images to reduce metal artifacts in kV-CT images.

[0088] FIG5 is an exemplary flowchart of determining a template image according to some embodiments of this specification.

[0089] Process 500 can be executed by a processing device (e.g., processing device 140). For example, process 500 can be implemented as an instruction set (e.g., an application) that is stored in a memory internal to or external to metal artifact correction system 100. The processing device can execute the instruction set and, when executing the instructions, can be configured to perform process 500. The operational diagram of process 500 presented below is illustrative. In some embodiments, the process can be completed using one or more additional operations not described and / or omitting one or more operations discussed below. In addition, the order of the operations of process 500 shown in Figure 5 and described below is not intended to be limiting. In some embodiments, process 500 can be used to implement step 330 in process 300.

[0090] Step 510 : Divide the first image into a plurality of sub-regions. In some embodiments, step 510 may be performed by the processing device 140 or the template image determination module 230 .

[0091] In some embodiments, the processing device 140 may divide the entire first image into N sub-regions, where the N sub-regions may be of equal or different sizes. Specifically, the processing device 140 may automatically perform the division in a preset manner. For example, the first image may be divided into N grids (including 9 grids, 16 grids, etc.).

[0092] In some embodiments, processing device 140 can divide the first image into multiple sub-regions based on the delineation information. Delineation can be performed manually. For example, a region of bone tissue that can be discerned by the naked eye can be delineated and set as a sub-region. Alternatively, a region with a distinct boundary can be divided into different sub-regions based on the boundaries in the image. Delineation based on delineation information provides a more refined result than segmentation using a preset method.

[0093] In some embodiments, the processing device 140 may automatically perform the segmentation by machine learning. For example, the processing device 140 may process the first image using a trained region segmentation model, inputting the first image into the region segmentation model to generate a plurality of segmented sub-regions.

[0094] Step 520 : Determine a high-density subregion, a medium-density subregion, and a low-density subregion within each subregion based on at least two local thresholds. In some embodiments, step 520 may be performed by the processing device 140 or the template image determination module 230 .

[0095] In some embodiments, the processing device 140 may set M local thresholds in each sub-region to further divide each sub-region into M+1 sub-regions of different densities. Wherein, M is a natural number greater than or equal to 2. In some embodiments, the value range of M may include 2 to 10. In some embodiments, the value range of M may include 2 to 4. For example, when M is 2, each sub-region may be further divided into 3 sub-regions of different densities, which may include a high-density sub-region, a medium-density sub-region, and a low-density sub-region. Wherein, the number of local thresholds in each sub-region may be the same or different.

[0096] In some embodiments, the processing device 140 may compare the pixel points in each sub-region with the M local thresholds in the sub-region, and divide the pixel points into one of the M+1 regions based on the comparison results, thereby determining the regional distribution within the sub-region. As an example only, the processing device 140 may first divide the first image into 9 sub-regions, and set two local thresholds T in the first sub-region. max and T min When the value of a pixel in the first sub-region is greater than T max When the value of a pixel in the first sub-region is less than or equal to T max and greater than or equal to T min When T max and T minIf the value of a pixel in the first sub-region is less than T min When , the pixel can be divided into the low-density sub-region. When each pixel in the first sub-region is judged, the division of the high-density sub-region, medium-density sub-region and low-density sub-region in the sub-region can be known.

[0097] In some embodiments, the processing device 140 may set the same or different local thresholds based on the conditions within each sub-region. For example, for sub-regions that are automatically divided by a preset method, the density distribution in each sub-region may be the same, and the same local threshold may be set. For sub-regions divided by delineation information, if there may be obvious metal areas in a sub-region, a higher local threshold may be set; if there is obviously no metal area in a sub-region and most of it is soft tissue, a lower local threshold may be set. For example only, the local threshold may be set to T max ' and T min ', and T max '>T min ', T max '>T max , T min '﹤T min .

[0098] In some embodiments, the local threshold value in each sub-region can be determined based on the maximum value and the minimum value of the pixels in each sub-region. For example, it can be determined based on the sum of the maximum value and the minimum value of the pixels multiplied by the threshold coefficient. For example, the local threshold value T max It can be set to (maximum value of pixel + minimum value of pixel) * first coefficient, local threshold T min It can be set as (maximum pixel value + minimum pixel value) * second coefficient, and the first coefficient is greater than the second coefficient. The threshold coefficients (e.g., first coefficient and second coefficient) can be set manually and can be adjusted based on the situation in the sub-region. In some embodiments, the processing device 140 can further use machine learning methods to determine the local threshold and / or threshold coefficient.

[0099] In some embodiments, different density areas generally need to include at least three areas, including a high-density area M1, a medium-density area M2, and a low-density area M3, to correspond to bone tissue and implanted metal areas, soft tissue areas and in vivo cavities, and in vitro air areas, respectively. Therefore, for each sub-area, at least two local thresholds can be set to divide the above three density areas, for distinguishing bone tissue, soft tissue, and air, etc. Furthermore, the processing device 140 can also set more than two local thresholds, for example, 3, 4, or 5 local thresholds, so that areas with more density levels can be divided, for example, areas with 4, 5, or 6 different density levels. The division of more density areas can have more refined segmentation results. The number of local thresholds is related to the number of different density areas. If M local thresholds are set, the number of different density areas is M+1.

[0100] In step 530 , the high-density sub-region, the medium-density sub-region, and the low-density sub-region in each sub-region are merged to generate a high-density region, a medium-density region, and a low-density region. In some embodiments, step 530 may be performed by the processing device 140 or the template image determination module 230 .

[0101] In some embodiments, after all pixel points in each sub-area have been judged, the processing device 140 can merge the high-density sub-areas in all sub-areas to generate a high-density area, merge the medium-density sub-areas in all sub-areas to generate a medium-density area, and merge the low-density sub-areas in all sub-areas to generate a low-density area.

[0102] Taking the above-mentioned 9 sub-regions as an example, when all the pixel points in the 9th sub-region are judged, the division of high-density sub-regions, medium-density sub-regions and low-density sub-regions in each sub-region can be known. By merging the high-density sub-regions in the 9 sub-regions, merging the medium-density sub-regions in the 9 sub-regions, and merging the low-density sub-regions in the 9 sub-regions, the division of high-density regions, medium-density regions and low-density regions of the entire first image can be obtained.

[0103] In some embodiments, processing device 140 can further generate bone and metal masks for the overall high-density regions, soft tissue masks for the medium-density regions, and air masks for the low-density regions. Since subsequent calculations only require the demarcation of regions with different densities, generating separate mask images is not necessary.

[0104] In some embodiments, the processing device 140 may first process the first image based on a local threshold segmentation method to determine high-density regions and medium-density regions, then fuse the second image with the first image to generate a composite image, and finally process the composite image based on the local threshold segmentation method to determine low-density regions. The specific process for determining regions of different densities is consistent with the above steps.

[0105] The first image (i.e., the MV-CBCT image) shown in FIG4A lacks the bedplate portion (the lowermost portion of FIG4B ) compared to the second image (i.e., the kV-CT image) shown in FIG4B . In some embodiments, the processing device 140 may integrate / copy the bedplate region information in the second image to the same location in the first image before determining the different density regions to generate a composite image. Alternatively, after generating the air mask, the processing device 140 may integrate / copy the bedplate region information in the air mask into the bone tissue, metal mask, and soft tissue mask. Integrating / copying the bedplate region information in the second image into the first image ensures data equivalence between the first and second images, thereby avoiding problems during subsequent image processing.

[0106] Step 540 , based on the CT values ​​of the regions corresponding to the different density regions in the second image, fill the different density regions to obtain a template image. In some embodiments, step 540 may be performed by the processing device 140 or the template image determination module 230 .

[0107] After the first and second images are registered in step 320, the pixel matrix size of the first image and the pixel matrix size of the second image are identical. Therefore, each pixel in the different density regions of the first image determined in step 530 has a unique corresponding pixel in the second image, thereby determining the different density regions in the second image corresponding to the first image. Specifically, processing device 140 can determine the different density regions of the second image based on the different density regions of the first image. Based on the density region to which each pixel in the first image belongs, processing device 140 can determine that the corresponding pixel in the second image also belongs to the corresponding density region. The different density regions of the second image may include a first region M1', a second region M2', and a third region M3'. The high-density region M1 of the first image corresponds to the first region M1' of the second image, the medium-density region M2 of the first image corresponds to the second region M2' of the second image, and the low-density region M3 of the first image corresponds to the third region M3' of the second image. By way of example only, if pixel Z in the first image belongs to the low-density region, the corresponding pixel Z' in the second image can be determined to belong to the third region M3'.

[0108] Because the CT values ​​of different tissues in the first and second images differ significantly, it may be considered to fill pixel values ​​in different density regions of the first image using the CT value of the second image. Specifically, the processing device 140 may fill the high-density region M1 based on the CT value of the first region M1' in the second image. The first region M1' is the region in the second image corresponding to the high-density region M1. Furthermore, the processing device 140 may fill the medium-density region M2 based on the average CT value of each pixel in the second region M2' in the second image. The second region M2' is the region in the second image corresponding to the medium-density region M2. Finally, the processing device 140 may fill the low-density region M3 based on the average CT value of each pixel in the third region M3' in the second image. The third region M3' is the region in the second image corresponding to the low-density region M3. In some embodiments, the processing device 140 may generate the template image based on the filled high-density region, medium-density region, and intermediate-density region.

[0109] Specifically, the processing device 140 may fill the different density regions using the following formula (2):

[0110]

[0111] Among them, I template represents the template image, I CT is the pixel matrix of the second image, (i, j) is the different pixel points in the image pixel matrix, N M2′ and N M3′ The number of pixels in the second area M2' and the third area M3' respectively. i and j, and N M2′ and N M3′ The value range of is related to the size of the image. For example, if the image is a 512*512*512 pixel matrix, the value range of i and j is 1 to 512, and N M2′ and N M3′ The value range is 1 to 512 3 If the image is a 1024*1024*1024 pixel matrix, the value range of i and j is 1 to 1024, N M2′ and N M3′ The value range is 1 to 1024 3 In the process of generating the template image, the second image is first used as the template image, and then the different divided areas are filled respectively using the above formula (2).

[0112] It should be noted that the above description of process 500 is for illustration and purpose only and does not limit the scope of application of this specification. Those skilled in the art may make various modifications and variations to process 500 under the guidance of this specification. However, such modifications and variations are still within the scope of this specification.

[0113] FIG6 is an exemplary flowchart of determining a template image according to some embodiments of the present specification.

[0114] As shown in Figure 6, the method for determining a template image described in an embodiment of this specification can first perform image registration on the first image (for example, an MV-CBCT image) and the second image (for example, an original CT image). Furthermore, different density areas of the first image are determined based on the registered first and second images. Specifically, the different density areas include a high-density area M1 (corresponding to bone tissue and metal mask), a medium-density area M2 (corresponding to soft tissue mask), and a low-density area M3 (corresponding to air mask). Among them, the first image and the second image can be first fused to generate a composite image (i.e., a CT / CBCT composite image), and then the low-density area M3 is determined based on the composite image. Finally, based on the different density areas (including the high-density area M1, the medium-density area M2, and the low-density area M1), the CT value of the second image is used to fill in to generate a template image.

[0115] FIG. 7 is an exemplary flowchart of determining a metal artifact image according to some embodiments of the present specification.

[0116] Process 700 can be executed by a processing device (e.g., processing device 140). For example, process 700 can be implemented as an instruction set (e.g., an application) that is stored in a memory internal to or external to metal artifact correction system 100. The processing device can execute the instruction set, and when executing the instructions, it can be configured to perform process 700. The operational diagram of process 700 presented below is illustrative. In some embodiments, the process can be completed using one or more additional operations not described and / or omitting one or more operations discussed below. In addition, the order of the operations of process 700 shown in Figure 7 and described below is not intended to be limiting. In some embodiments, process 700 can be used to implement step 340 in process 300.

[0117] In this embodiment, since the second image is primarily affected by metal in the same ray path as the metal, to correct for the metal's effects on medium- and low-density areas, it is necessary to first determine the metal projection area. Based on the difference in projections between the second image and the template image in the metal projection area, metal artifacts in non-high-density areas are obtained for subsequent correction. The steps for determining the metal projection area and performing correction are described in detail below:

[0118] Step 710 : Determine a metal region based on the second image. In some embodiments, step 710 may be performed by the processing device 140 or the metal artifact determination module 240 .

[0119] The metal region is a region in the second image that contains only metal parts (e.g., prosthetic joints, dental fillings, spinal fixation bolts, etc.), without bone tissue or artifacts. Processing device 140 can extract the metal region from the first region M1' or the bone tissue and metal mask. Although the presence of metal artifacts can affect the quality of the reconstructed image and impact subsequent image applications, it is important to retain information about the actual location of the metal implant within the target object during image metal artifact correction.

[0120] In some embodiments, the processing device 140 may determine the metal region M by using a threshold segmentation method. metal Since the second image is a kV-CT image, the CT values ​​corresponding to metal in the second image are very high, possibly even reaching tens of thousands of Hounsfield units. Therefore, it is possible to consider setting a metal threshold to screen out bone tissue and metal artifacts in the first region M1'. In contrast, since the first image is obtained by a megavoltage CT device, the difference in CT values ​​between the metal part and the bone tissue is not so obvious, making it difficult to distinguish. Therefore, the second image needs to be processed separately, and the metal part cannot be directly separated when determining different density areas.

[0121] Specifically, the metal area M is determined by segmentation method. metal When the processing device 140 compares each pixel in the first region M1' of the first image or the bone tissue and metal mask portion with a preset metal threshold, and determines whether the pixel belongs to the metal region based on the comparison result. As an example only, if a pixel is greater than or equal to the metal threshold, the pixel can be determined to belong to the metal region and retained; if a pixel is less than the metal threshold, the pixel can be determined not to belong to the metal region and filtered out. After completing the comparison of all pixels in the first region M1', all retained pixels can constitute the metal region.

[0122] In some embodiments, the metal threshold can be set manually or adjusted according to the situation in the first image. For example, the metal threshold can be directly set to 10,000 Hu.

[0123] Step 720 : Perform a front projection on the metal area to generate a metal projection map. The non-zero area in the metal projection map is the metal projection area. In some embodiments, step 720 may be performed by the processing device 140 or the metal artifact determination module 240 .

[0124] The metal projection image is a projection sinusoidal graph obtained by front-projecting the metal area. A metal projection sinusoidal graph is an image that displays projection data using a sinusoidal graph, representing projection data obtained at different projection angles. Combining the non-zero regions in the metal projection sinusoidal graph represents the metal projection area in the projection domain. In some embodiments, the horizontal axis of the metal projection image can represent the projection angle of the detector, and the vertical axis can represent the pixel position of the detector.

[0125] In some embodiments, as described in step 350 , during the iterative correction process, since the position of the metal implant is fixed, there is no need to repeatedly determine the metal area and the metal projection area based on the updated second image.

[0126] Step 730 : Perform a front projection on the second image to generate a second image projection map. In some embodiments, step 730 may be performed by the processing device 140 or the metal artifact determination module 240 .

[0127] In some embodiments, the processing device 140 may also perform forward projection on the complete second image (e.g., kV-CT image) to generate a second image projection map, so as to process the image data in the projection domain. The second image projection map is a projection sinogram obtained by projecting the complete second image.

[0128] In some embodiments, steps 710, 720, and 730 are performed on the second image and are not related to the step of determining the template image. Therefore, steps 710, 720, and 730 can be performed after the first image and the second image are registered in step 320. They do not need to be performed after the step of determining the template image in step 330, but only need to be completed before step 750. In some embodiments, steps 710, 720, 730 and step 330 can be performed simultaneously.

[0129] Step 740 , performing front projection on the template image to generate a template image projection image. In some embodiments, step 740 may be performed by the processing device 140 or the metal artifact determination module 240 .

[0130] In some embodiments, the processing device 140 may further perform a forward projection on the template image obtained in step 330 to generate a template image projection sinusoidal graph for processing the template image projection graph in the projection domain. The template image projection graph is a projection sinusoidal graph obtained after forward projection of the template image.

[0131] In some embodiments, step 740 and steps 710 , 720 , and 730 may be performed simultaneously, but must be performed after step 330 is completed.

[0132] Step 750 : Determine the difference between the projection image of the second image and the projection image of the template image in the metal projection area. In some embodiments, step 750 may be performed by the processing device 140 or the metal artifact determination module 240 .

[0133] Since the CT values ​​of different areas in the template image are obtained by filling the CT values ​​or average CT values ​​of the corresponding areas in the second image (for example, kV-CT image), when the template image and the second image are forward projected, the difference in the integral values ​​on the same beam path between the template image and the second image mainly comes from metal artifacts, and the formation of metal artifacts is caused by the signal deviation on the beam path passing through the metal. Among them, the same beam path can be understood as the same area, that is, the same density area. The difference in integral values ​​is the difference in integral values ​​of ray attenuation during the projection process, which is mainly used to express the difference in grayscale between the two images. Therefore, as long as the difference in sinusoidal diagrams of the metal projection area is extracted, the second image can be corrected. It can be seen from formula (2) that for the template image and the second image, the CT values ​​of their high-density areas are the same, while the CT values ​​of the medium-density area and the low-density area are different. By determining the difference in projection images, the metal artifacts in the medium-density area and the low-density area can be determined.

[0134] In some embodiments, the processing device 140 may determine the projection image difference based on the second image projection image, the template image projection image, and the metal projection area determined in step 720. Specifically, the projection image difference may be obtained based on the following formula (3):

[0135] P MetalDiff (θ,k)=FP(I CT )-FP(I template ),(θ,k)∈FP(M metal )≠0 (3)

[0136] Among them, P MetalDiff represents the projection difference, FP is the front projection operator, θ is the projection angle of the detector, and k is the detector pixel position. Wherein, the value range of θ may include 0 to 360°, and the value range of k is related to the resolution of the detector. In some embodiments, the value range of k may include 256 to 2048. FP(I CT ) can represent the second image projection, FP(I template ) can represent the template image projection, FP(M metal ) can represent the metal projection diagram, FP(M metal )≠0, it can represent the non-zero area in the metal projection map, that is, the metal projection area. By directly subtracting the values ​​of the second image projection map and the template image projection map in the metal projection area, the projection map difference P between the second image projection map and the template image projection map in the metal projection area can be obtained.MetalDiff .

[0137] Step 760 : Determine a metal artifact image based on the projection image difference. In some embodiments, step 760 may be performed by the processing device 140 or the metal artifact determination module 240 .

[0138] In some embodiments, the processing device 140 may convert the projection image difference P obtained in step 750 into MetalDiff Back-projection reconstruction is directly performed to determine the metal artifact image, and the determined metal artifact image is then substituted into step 350 to correct the second image to obtain a corrected image.

[0139] However, when performing CT imaging on a normal, metal-free target object, it's unlikely that a sharp drop in CT values ​​will occur due to metal regions; instead, the image will show a relatively smooth downward curve. Therefore, due to the presence of metal regions, projection values ​​may be truncated at the boundary regions of the projection image difference. Without smoothing, this could introduce new streak artifacts during subsequent processing. In this embodiment, smoothing the boundary regions of the projection image difference before determining the metal artifact image is considered, so that the projection values ​​at the boundary regions gradually decrease to zero, thereby avoiding the introduction of new streak artifacts that could degrade image quality.

[0140] In some embodiments, the processing device 140 may smooth the boundary region of the projection image difference to obtain a processed projection image difference. Furthermore, the processing device 140 may perform back-projection reconstruction based on the processed projection image difference to obtain a metal artifact image. In some embodiments, the processing device 140 may smooth the boundary region using linear interpolation or normalized interpolation. In some embodiments, the processing device 140 may determine the extended distances of the upper and lower boundaries of the boundary region of the projection value difference and perform linear interpolation on the extended distances to obtain the processed projection image difference.

[0141] Specifically, linear interpolation can be performed based on the following formula (4) to obtain the processed projection image difference:

[0142]

[0143] Among them, P interp represents the difference in projection images after interpolation, k n1 and k n2 are the upper and lower boundaries of a certain metal area under a certain projection angle, Δ n1 , Δ n2 are the upper and lower boundary expansion distances respectively. k n1 and k n2 The value range of Δn1 , Δ n2 The value range may include 1 to 100. Among them, the expansion distance of the upper and lower boundaries can be set manually or adjusted according to the correction results. In some embodiments, the processing device 140 can also determine the expansion distance of the upper and lower boundaries based on a machine learning model. In some embodiments, the machine learning model may include but is not limited to a combination of one or more of a neural network model, a convolutional neural network model, a visual geometry group network model, a full-resolution residual network model, a mask area convolutional neural network model, a multi-dimensional recurrent neural network model, etc. In some embodiments, the processing device 140 may input the projection map difference into the machine learning model to output the expansion distance of the upper and lower boundaries. The machine learning model can be trained based on a large number of projection map difference samples with labels (for example, correction results).

[0144] In some embodiments, the processed projection image difference is back-projected and reconstructed to obtain a metal artifact image, namely, a metal artifact image I metal It can be expressed as I metal =BP(P interp ), BP is the back-projection reconstruction operator. Using the smoothed projection image to determine the metal artifact image can effectively avoid the introduction of new artifacts due to projection value truncation.

[0145] It should be noted that the above description of process 700 is for illustration and purpose only and does 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 process 700. However, such modifications and alterations remain within the scope of this specification. For example, steps 710, 720, 730, and 740 may be performed simultaneously.

[0146] FIG8 is an exemplary flowchart of metal artifact correction according to some embodiments of the present specification.

[0147] As shown in FIG8 , the metal artifact correction method described in an embodiment of this specification can first perform image registration on a first image (e.g., an MV-CBCT image) and a second image (e.g., an original CT image). Furthermore, a template image can be determined based on the registered first and second images. The obtained template image and second image can then be forward projected to generate a template image projection map and a second image projection map (i.e., the original image projection map). At the same time, a metal region can be extracted from the second image, forward projected onto the metal region, and the metal projection region can be determined. Next, the projection map difference between the second image projection map and the template image projection map in the metal projection region can be determined based on the template image projection map and the second image projection map. After smoothing the projection region boundary of the projection map difference, the projection map difference is reconstructed through back-projection to generate a metal artifact image. The obtained metal artifact image can then be further subjected to high-frequency filtering, and finally, the processed metal artifact image can be subtracted from the second image to obtain a corrected image. After completing one round of correction, the corrected image can be updated to a new second image for one or more rounds of iteration. Specifically, an updated template image and an updated second image projection map may be generated based on the corrected image.

[0148] FIG. 9 is a diagram comparing the effects of metal artifact correction methods according to some embodiments of this specification.

[0149] In order to further understand the effect of the metal artifact correction method described in the embodiment of this specification, a simulation phantom and a real human body are considered for verification. During the verification process, a metal ball is placed inside the simulation phantom, and a metal denture is implanted in the real human body.

[0150] As shown in Figure 9, group (a) is the data of the simulation phantom, group (b) is the data of the real human body 1, and groups (c) and (d) are the data of different layers of the real human body 2. As can be seen from the first column of original CT images, the original CT images of the simulation phantom have metal artifacts due to the placement of metal balls. The real human bodies 1 and 2 also have metal artifacts due to the implantation of metal dentures. The second, third, and fourth columns are the images obtained after processing the same layer of data using the linear interpolation method (LIMAR), the prior image normalization interpolation method (NMAR), and the metal artifact correction method described in the embodiments of this specification, and all images have been adjusted to the same window width and window level. As can be seen from Figure 9, although the metal artifacts in the images processed by the linear interpolation method and the normalization method have been weakened compared to the original CT images, the metal artifacts are still relatively obvious from the overall image, which may affect subsequent applications. In the image data processed by the metal artifact correction method described in the embodiments of this specification, the metal artifacts have been greatly weakened and it is almost impossible to clearly see that there are still metal artifacts.

[0151] Specifically, on the one hand, the linear interpolation method can obtain the projection of the metal area by linearly interpolating the projection values ​​on both sides of the metal area projection. The linear interpolation method reduces the artifacts caused by the back-projection reconstruction of the metal area to a certain extent, but the interpolation area will cause the loss of projection information. For the artifacts caused by a single relatively small metal implant, it can achieve a good repair effect, but for the case of multiple metals or large metal implants, the use of linear interpolation will cause too much projection information loss, and new artifacts will be introduced into the reconstructed image or some details will be lost. On the other hand, the method of using prior image normalization interpolation can use the original image to obtain a prior image, normalize the projection value of the original image by the projection value of the prior image, and then interpolate the metal area. This normalized interpolation method reduces the discontinuity of the metal area interpolation and has a better image repair effect than the linear interpolation method. However, when the metal artifact is more serious, it is difficult to obtain an ideal prior image, and the image repair effect is limited. In contrast, the metal artifact correction method described in the embodiments of this specification registers a first image at a first energy level with a second image at a second energy level. A template image is then generated using the geometric information of different tissue regions in the first image and the CT values ​​in the second image, ultimately achieving image restoration. This method achieves excellent restoration results even in the presence of multiple or large metals, restoring image geometry while maintaining the accuracy of CT values, meeting the requirements of clinical target delineation and dose calculation.

[0152] Further analysis of the image data was performed by selecting different regions A and B in each image set and calculating the CT values ​​and mean squared deviations within these regions. As shown in Figure 9, regions A and B in each image set are consistent with the original CT images. The average CT values ​​and mean squared deviations for regions A and B in the different data sets are shown in Table 1 below:

[0153] Table 1 Average CT values ​​and mean square error of different groups of data in different areas

[0154]

[0155]

[0156] The reference value of group a (simulation phantom) data is the CT value in the absence of metal. Due to the influence of metal artifacts, the reference values ​​of groups (b) to (d) cannot be accurately obtained. Therefore, the CT values ​​of the surrounding artifact-free areas are used instead to provide an approximate range of CT values.

[0157] The data in Table 1 above demonstrates that different correction methods can, to a certain extent, improve the accuracy of image CT values ​​and reduce CT value nonuniformity. Comparing statistical results, the metal artifact correction method described in the embodiments of this specification achieves superior correction results compared to other correction methods, with the CT values ​​of the corrected images being closer to the reference CT values.

[0158] The beneficial effects that may be brought about by the embodiments of this specification include but are not limited to: (1) in the process of determining different density areas, the overall image is first divided into multiple sub-areas, and then a local threshold segmentation method is performed, which can make the final results of high-density area, medium-density area and low-density area more accurate; (2) the boundary area where the projection image of the second image and the template image projection image differ in the metal projection area is smoothed, which can effectively avoid the introduction of new artifacts due to projection value truncation; (3) the obtained metal artifact image is filtered, which can further reduce the errors caused by the parameter selection of other steps; (4) by updating the corrected image to the second image for iterative correction, the artifacts can be gradually removed and the image correction effect can be improved.

[0159] 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.

[0160] 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.

[0161] 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.

[0162] 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.

[0163] 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.

[0164] 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.

[0165] 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.

[0166] 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 metal artifact correction method, characterized in that: include: Acquire a first image obtained by a first imaging device and a second image obtained by a second imaging device, wherein the second imaging device is a CT device; registering the first image and the second image; determining a template image based on the registered first image and the second image; determining a metal artifact image based on the template image; as well as The second image is corrected based on the metal artifact image to obtain a corrected image.

2. The method according to claim 1, characterized in that The first imaging device is a first energy level CT device, the second imaging device is a second energy level CT device, and the first energy level is greater than the second energy level.

3. The method according to claim 1 or 2, characterized in that The first imaging device is a megavoltage CT device, and the second imaging device is a kilovoltage CT device.

4. The method according to any one of claims 1 to 3, characterized in that The first image is a cone-beam CT image, and the second image is a diagnostic-level CT image.

5. The method according to any one of claims 1 to 4, characterized in that The determining of the template image based on the registered first image and the second image includes: determining regions of different densities based on the first image; The different density regions are filled based on CT values ​​of regions corresponding to the different density regions in the second image to obtain the template image.

6. The method according to claim 5, characterized in that The different density areas include a high-density area, a medium-density area and a low-density area.

7. The method according to claim 6, characterized in that Determining different density areas based on the first image includes: The first image is processed based on a local threshold segmentation method to determine the high-density area, the medium-density area, and the low-density area.

8. The method according to claim 7, characterized in that The processing of the first image based on the local threshold segmentation method to determine the high-density area, the medium-density area, and the low-density area includes: dividing the first image into a plurality of sub-regions; determining a high-density sub-region, a medium-density sub-region, and a low-density sub-region within each sub-region based on at least two local thresholds; and The high-density sub-region, the medium-density sub-region and the low-density sub-region in each sub-region are merged to generate the high-density region, the medium-density region and the low-density region respectively.

9. The method according to claim 8, characterized in that The dividing the first image into a plurality of sub-regions comprises: The first image is divided into the plurality of sub-regions based on delineation information.

10. The method according to any one of claims 6 to 9, characterized in that Determining different density areas based on the first image includes: Processing the first image based on a local threshold segmentation method to determine the high-density area and the medium-density area; fusing the second image with the first image to generate a composite image; and The composite image is processed based on the local threshold segmentation method to determine the low-density area.

11. The method according to any one of claims 6 to 10, characterized in that Filling the different density regions based on the CT values ​​of the regions corresponding to the different density regions in the second image to obtain the template image includes: filling the high-density area based on a CT value of a first area in the second image, where the first area is an area in the second image corresponding to the high-density area; filling the medium-density area based on an average CT value of each pixel point in a second area of ​​the second image, where the second area is an area in the second image corresponding to the medium-density area; Filling the low-density area based on an average CT value of each pixel point in a third area in the second image, the third area being an area in the second image corresponding to the low-density area; and The template image is generated based on the filled high-density area, medium-density area and medium-density area.

12. The method according to any one of claims 1 to 11, characterized in that Determining the metal artifact image based on the template image includes: determining a metal area based on the second image; Performing a front projection on the metal area to generate a metal projection map, wherein a non-zero area in the metal projection map is the metal projection area; performing front projection on the second image to generate a second image projection diagram; Performing forward projection on the template image to generate a template image projection image; Determining a difference between the second image projection and the template image projection in the metal projection area; and The metal artifact image is determined based on the projection image difference.

13. The method according to claim 12, characterized in that Determining the metal artifact image based on the projection image difference includes: performing smoothing processing on the boundary area of ​​the projection image difference to obtain a processed projection image difference; and Based on the processed projection image differences, back-projection reconstruction is performed to obtain a metal artifact image.

14. The method according to claim 13, characterized in that The smoothing of the boundary area in the projection image difference comprises: Determining the extension distances of the upper and lower boundaries of the boundary area; and Linear interpolation processing is performed based on the extended distances of the upper and lower boundaries to obtain a processed projection image difference.

15. The method according to claim 14, characterized in that Determining the extension distances of the upper and lower boundaries of the boundary area includes: The extension distance of the upper and lower boundaries is determined based on a machine learning model.

16. The method according to any one of claims 1 to 15, characterized in that Correcting the second image based on the metal artifact image to obtain a corrected image includes: performing filtering processing on the metal artifact image to obtain a processed metal artifact image; and The second image is corrected based on the processed metal artifact image to obtain the corrected image.

17. The method according to any one of claims 1 to 16, characterized in that Correcting the second image based on the metal artifact image to obtain a corrected image includes: The metal artifact image is subtracted from the second image to obtain the corrected image.

18. The method according to any one of claims 1 to 17, characterized in that The method further comprises: Based on the corrected image, the second image is updated for one or more iterations.

19. The method according to claim 18, characterized in that Based on the corrected image, updating the second image for one or more iterations includes: updating the template image based on the updated second image; updating the metal artifact image based on the updated template image; The updated second image is corrected based on the updated metal artifact image to obtain an updated corrected image.

20. A metal artifact correction system, characterized in that: include: an acquisition module, configured to acquire a first image obtained by a first imaging device and a second image obtained by a second imaging device, wherein the second imaging device is a CT device; a registration module, configured to register the first image and the second image; a template image determining module, configured to determine a template image based on the registered first image and the second image; a metal artifact determination module, configured to determine a metal artifact image based on the template image; as well as A correction module is used to correct the second image based on the metal artifact image to obtain a corrected image.

21. A metal artifact correction system, characterized in that: The system comprises: at least one storage device for storing computer instructions; and at least one processor configured to communicate with the at least one memory device, wherein when executing the set of instructions, the at least one processor is configured to: Acquire a first image obtained by a first imaging device and a second image obtained by a second imaging device, wherein the second imaging device is a CT device; registering the first image and the second image; determining a template image based on the registered first image and the second image; determining a metal artifact image based on the template image; and The second image is corrected based on the metal artifact image to obtain a corrected image.

22. A metal artifact correction device, characterized in that: The device includes at least one processor and at least one memory, the at least one memory is used to store computer instructions, and the at least one processor is used to execute at least part of the computer instructions to implement the metal artifact correction method according to any one of claims 1 to 19.

23. A computer-readable storage medium, characterized in that The storage medium stores computer instructions, and when the computer instructions are executed by a processor, the metal artifact correction method according to any one of claims 1 to 19 is implemented.

Images