Image correction method, system and related device for SPECT (Single Photon Emission Computed Tomography) image

Abstract

The invention discloses an image correction method and system for an SPECT image and a related device, and relates to the field of image processing, and the method comprises the steps: obtaining an SPECT image and a CT image of the same patient, carrying out the alignment, extracting the function distribution characteristics of the SPECT image and the structural characteristics of the CT image through a preset deep learning model, and determining the artifact probability value of a pixel point in the SPECT image; classifying pixel points of the SPECT image through an artifact probability threshold, and generating a binary mask graph of the SPECT image according to a classification result; determining an artifact area of the SPECT image through the binary mask image, and correcting a pixel value of the artifact area through structural information of the CT image; and fusing the corrected artifact region with the non-artifact region of the SPECT image to obtain a corrected SPECT image. According to the method, the SPECT image is corrected by using the structural information of the CT image, and the correction accuracy of the SPECT image is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, in particular to a SPECT image correction method, system and related device. BACKGROUND

[0002] SPECT (Single Photon Emission Computed Tomography) is a nuclear medical imaging technology, and its principle is: radioactive drugs are injected into the patient's body, and the radioactive drugs can release single photons (gamma rays) in different tissues or different organs in the human body. The detector receives the signal released by the radioactive drug to obtain the original projection data, and the computer reconstructs the tomographic image (SPECT image) of the tissue or organ in the human body according to the original projection data. The reconstruction refers to the process of converting the original projection data collected by the detector into a tomographic image through a mathematical algorithm.

[0003] Because the radioactive drug can accumulate in a specific tissue or organ, and the distribution of the radioactive drug is related to the blood, metabolism, and nerve conduction functions of the tissue or organ, the SPECT image can include the functional state of the tissue or organ (such as the blood distribution state, metabolic state, etc. of the tissue or organ).

[0004] Due to factors such as scanning device defects and acquisition time, part of the projection data is missing or incomplete, which often causes false images in the SPECT image that do not belong to the human tissue or organ, and the false images can affect the accuracy of diagnosis. At present, the SPECT image is mainly corrected for false images by optimizing the algorithm, which calculates the missing projection data from the existing projection data and corrects the false images of the SPECT image by supplementing the SPECT image.

[0005] Because the optimization algorithm often relies on prior assumptions to supplement the SPECT image to correct the SPECT image, if the prior assumptions do not conform to reality, the supplemented SPECT image will be distorted, resulting in low correction accuracy of the SPECT image. SUMMARY

[0006] In view of the above problems, the present application provides a SPECT image correction method, system and related device to improve the correction accuracy of the SPECT image. The specific scheme is as follows:

[0007] The first aspect of the present application provides a SPECT image correction method, which comprises:

[0008] acquire a single photon emission computed tomography (SPECT) image and a computed tomography (CT) image of a same patient, and align a coordinate system of the SPECT image and a coordinate system of the CT image so that a pixel point of the SPECT image and a pixel point of the CT image have a corresponding relationship;

[0009] extract a functional distribution feature of the SPECT image and a structural feature of the CT image by a preset deep learning model, and determine a artifact probability value of each pixel point in the SPECT image according to the functional distribution feature and the structural feature;

[0010] classify the pixel points in the SPECT image into an artifact category and a non-artifact category by comparing an artifact probability threshold value and the artifact probability value of each pixel point, and generate a binary mask image of the SPECT image according to a classification result of the pixel points;

[0011] determine at least one artifact region in the SPECT image by the binary mask image, and correct a pixel value of the artifact region by structural information of the CT image in the artifact region;

[0012] fuse the corrected artifact region and a non-artifact region of the SPECT image to obtain a corrected SPECT image.

[0013] In a possible implementation, the acquiring a single photon emission computed tomography (SPECT) image and a computed tomography (CT) image of a same patient comprises:

[0014] The SPECT image and the CT image of the same patient are simultaneously acquired by a SPECT / CT integrated machine.

[0015] In a possible implementation, the training data of the preset deep learning model comprises: a plurality of groups of SPECT images and CT images paired with the SPECT images, each group of SPECT images is labeled with a plurality of functional regions, and each group of CT images is labeled with a plurality of structural regions.

[0016] In a possible implementation, the classifying the pixel points in the SPECT image into an artifact category and a non-artifact category by comparing an artifact probability threshold value and the artifact probability value of each pixel point comprises:

[0017] For each pixel point in the SPECT image, if the artifact probability value of the pixel point is greater than the artifact probability threshold value, the category of the pixel point is the artifact category, and if the artifact probability value of the pixel point is not greater than the artifact probability threshold value, the category of the pixel point is the non-artifact category;

[0018] Obtain the classification result of the pixel point in the SPECT image.

[0019] In a possible implementation, the mask pixel points of the binary mask image correspond to the pixel points of the SPECT image one by one.

[0020] The generating of the binary mask image of the SPECT image according to the classification result of the pixel point comprises:

[0021] Traversing each pixel point in the SPECT image, if the category of the pixel point is the artifact category, the pixel value of the mask pixel point corresponding to the pixel point is set to 1;

[0022] If the category of the pixel point is the non-artifact category, the pixel value of the mask pixel point corresponding to the pixel point is set to 0.

[0023] Obtain the binary mask image.

[0024] In a possible implementation, the correcting of the pixel value of the artifact region through the structural information of the CT image comprises:

[0025] The correcting of the gray value of the artifact region through the structural information of the CT image and the image correlation relationship, the image correlation relationship being a feature correlation relationship between a functional region in the SPECT image and a structural region in the CT image.

[0026] In a possible implementation, the method further comprises:

[0027] Evaluating the image quality of the corrected SPECT image, if the image quality does not meet the requirement, performing secondary iteration correction on the corrected SPECT image.

[0028] In a possible implementation, the method further comprises:

[0029] In the reconstruction process of the SPECT image, the structural information of the CT image is taken as a constraint condition of the reconstruction process.

[0030] The second aspect of the present application provides a SPECT image correction system, the SPECT image correction system comprising:

[0031] An acquisition unit is configured to acquire a single photon emission computed tomography (SPECT) image and a computed tomography (CT) image of a same patient, and align a coordinate system of the SPECT image and a coordinate system of the CT image, so that a pixel point of the SPECT image and a pixel point of the CT image have a corresponding relationship.

[0032] an analysis unit configured to extract functional distribution features of the SPECT image and structural features of the CT image by using a preset deep learning model, and determine a probability value of each pixel point in the SPECT image according to the functional distribution features and the structural features;

[0033] a classification unit configured to classify the pixel points in the SPECT image into a pseudo artifact category and a non-pseudo artifact category by comparing a pseudo artifact probability threshold value with the probability value of each pixel point, and generate a binary mask image of the SPECT image according to the classification result of the pixel points;

[0034] a correction unit configured to determine at least one pseudo artifact region in the SPECT image by using the binary mask image, and correct pixel values in the pseudo artifact region by using structural information of the CT image;

[0035] a fusion unit configured to fuse the corrected pseudo artifact region with a non-pseudo artifact region of the SPECT image, and obtain a corrected SPECT image.

[0036] The third aspect of the present application provides an electronic device, comprising at least one processor and a memory connected with the processor, wherein:

[0037] The memory is configured to store a computer program;

[0038] The processor is configured to execute the computer program, so that the electronic device can implement the SPECT image correction method of the first aspect or any implementation manner of the first aspect.

[0039] By means of the technical scheme, the application provides a SPECT image correction method, system and related device. The method acquires a SPECT image and a CT image of the same patient, and aligns the SPECT image and the CT image, so that the pixel points of the SPECT image and the pixel points of the CT image have a corresponding relationship. The functional distribution features of the SPECT image and the structural features of the CT image are extracted, and the artifact probability value of each pixel point in the SPECT image is determined according to the distribution features and the structural features. Each pixel point in the SPECT image is divided into an artifact category and a non-artifact category through an artifact probability threshold, and a binary mask image of the SPECT image is generated according to the classification result of the pixel points. At least one artifact region in the SPECT image is determined through the binary mask image, and the pixel value in the artifact region is corrected through the structural information of the CT image. The corrected artifact region and the non-artifact region of the SPECT image are fused to obtain a corrected SPECT image. The method learns and utilizes the image information correlation relationship between the CT image and the SPECT image, assists in identifying the artifact region in the SPECT image, and corrects the artifact region. The actual structural information of the CT image is used to correct the SPECT image, the dependence on prior assumptions is eliminated, and the correction accuracy of the SPECT image is effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0040] The above and other features, advantages, and aspects of the present disclosure will become more apparent by describing in detail exemplary embodiments thereof with reference to the attached drawings in which:

[0041] Figure 1 A flowchart of a SPECT image correction method provided by an embodiment of the application;

[0042] Figure 2 A structural diagram of a SPECT image correction system provided by an embodiment of the application;

[0043] Figure 3 A hardware structural diagram of an electronic device provided by the application. DETAILED DESCRIPTION

[0044] The embodiments of the application will be described below with reference to the accompanying drawings. The terms used in the embodiment part of the application are only used to explain the specific embodiments of the application, and are not intended to limit the application.

[0045] The embodiments of the present application will be described below in conjunction with the drawings. Those skilled in the art can know that, with the development of technology and the emergence of new scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0046] The terms "comprising" and "having" and any variations thereof in the specification and above drawings of the present application are intended to cover the non-exclusive inclusion, so that the processes, methods, systems, products or equipment containing a series of units do not have to be limited to those units, but can include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0047] In single photon emission computed tomography (SPECT), due to the influence of factors such as scanning equipment (such as equipment noise), acquisition time or patient motion, part of the projection data is missing or incomplete, resulting in artifacts in the SPECT image, or when the computer converts the original projection data, the reconstruction algorithm of the computer has limited ability to supplement and correct the missing data in the original projection data, resulting in distortion of the generated image and artifacts. Artifacts are not real structures or signals in the human body, but "false images" in the image. Artifacts can mask patient lesions or create false lesions, thereby affecting patient diagnosis.

[0048] Current SPECT image correction relies on interpolation methods or optimization algorithms, but both have the problem of poor correction effect. For example, optimization algorithms often rely on prior assumptions to supplement SPECT images, but if the prior assumptions do not conform to the century, they may instead cause SPECT image distortion, and optimization algorithms also rely on parameter selection, so if the parameter selection is not appropriate, it will also affect the correction effect of the SPECT image.

[0049] To solve the above problems, the embodiments of the present application provide a SPECT image correction method, which uses the structural information of the CT image to correct the artifacts of the SPECT image, eliminates the dependence on prior assumptions or parameters, and effectively improves the correction accuracy of the SPECT image. CT (Computed Tomography, computed tomography) is a medical imaging technology that rotates X-rays around the patient, acquires X-ray attenuation data from multiple angles through the detector, and converts the X-ray attenuation data into a tomographic image of the human body tissue or organ using a computer.

[0050] The CT image can include structural information of internal tissues or organs of a human body, such as shape, position, size, density and the like of the tissues or organs, reflecting the anatomical structure of the human body. Therefore, in the embodiment, the CT image can provide high-resolution anatomical structure information, especially in soft tissue and bone parts, while the SPECT image can reflect the functional state of the tissues or organs (such as blood distribution state, metabolic state and the like of the tissues or organs). Since each tissue or organ has its own structure and functional distribution fixed and can have a correlation, the CT image and the SPECT image have a certain image correlation, and by comparing the tissue distribution characteristics of the CT image and the SPECT image at the same position (such as the same tissue or the same organ), the region in the SPECT image that does not conform to the anatomical structure in the CT image can be identified, and this region is most likely to be an artifact region in the SPECT image. For example, for a normal heart, the internal anatomical structure is fixed, and the blood distribution in the heart is also fixed, and the structural information of the heart can have a correlation with the blood distribution, and if the structural information of the heart in the CT image does not correspond to the blood distribution in the heart region in the SPECT image (there is a difference between the original blood distribution corresponding to the structural information), it can be indicated that there is an artifact in the heart region of the SPECT image, affecting the correlation between the CT image and the SPECT image. Therefore, the structural information of the CT image can be used as reference data for the SPECT image to help identify and correct the artifact of the SPECT image. The SPECT image image correction method of the embodiment of the application will be described in detail below with reference to the accompanying drawings.

[0051] Referring to Figure 1 , Figure 1 A flowchart of a SPECT image image correction method provided by the embodiment of the application is shown in FIG. 1. The SPECT image image correction method provided by the embodiment of the application can include steps S10 to S14, which will be described in detail below. Figure 1

[0052] S10, obtaining a single photon emission computed tomography (SPECT) image and a computed tomography (CT) image of the same patient, and aligning the coordinate system of the SPECT image and the coordinate system of the CT image so that the pixel points of the SPECT image and the pixel points of the CT image have a corresponding relationship.

[0053] ​The SPECT image can be a tomographic image obtained by reconstruction of raw projection data by a computer, or can be a SPECT image generated by taking the structural information of the CT image as a constraint condition in the SPECT image reconstruction process of the computer. Specifically, in the reconstruction process of the SPECT image, the spatial matching relationship between the projection data of the SPECT image and the structural data of the CT image is jointly optimized, so that the functional information of the SPECT image is maintained, the structural information of the CT image is referred to, and the generation of artifacts of the SPECT image is reduced. Alternatively, in the reconstruction process of the SPECT image, the structural information of the CT image is introduced as a regularization constraint to constrain the reconstruction result of the SPECT image to be consistent with the structural information of the CT image, and the reconstruction process of the SPECT image is guided by the structural information of the CT image to reduce the generation of artifacts of the SPECT image. The present embodiment constrains the reconstruction process of the SPECT image by the structural information of the CT image, first reduces the generation of artifacts, and then corrects the artifacts through the artifact identification and repair network of the preset deep learning model, which can effectively improve the image quality of the SPECT image.

[0054] To improve the correction accuracy of the SPECT image, the present embodiment selects to obtain the SPECT image and the CT image of the same patient by a SPECT / CT integrated machine. The SPECT / CT integrated machine can provide two kinds of imaging data at the same time, so that the obtained CT image and SPECT image can have the same size and resolution. The SPECT / CT integrated machine is a medical imaging device that combines single photon emission computed tomography (SPECT) and computed tomography (CT). Of course, in another alternative embodiment, the CT image and the SPECT image can also be obtained from different data sources respectively.

[0055] When the CT image and the SPECT image are acquired, the CT image and the SPECT image are preprocessed and aligned in a coordinate system in this embodiment, so as to improve the quality of the image data and ensure the accurate alignment of the CT image and the SPECT image in the image region, thereby facilitating the determination of the corresponding relationship between the CT image and the SPECT image in the image region. The alignment of the coordinate system of the image can refer to adjusting the coordinate system of two or more images through a series of geometric transformations (such as translation, rotation, etc.), so that the spatial positions (pixel coordinates of a two-dimensional image) of the two or more images have a corresponding relationship. Specifically, in this embodiment, the purpose of the alignment of the coordinate system is to establish the corresponding relationship between the pixel points of the CT image and the SPECT image, so as to realize the correspondence of the image region. The alignment algorithm actually used in this embodiment can be a rigid registration algorithm (in the registration process, the transformation between images only includes translation and rotation, and does not involve any deformation (such as scaling, etc.)) or a non-rigid registration algorithm (in the registration process, the transformation between images includes not only translation and rotation, but also deformation).

[0056] Since the CT image and the SPECT image are acquired simultaneously from the SPECT / CT integrated machine, the image resolutions of the CT image and the SPECT image are the same, and a simple alignment operation can be performed through the rigid registration algorithm, so that the pixel points of the CT image and the pixel points of the SPECT image have a one-to-one corresponding relationship, and the two pixel points having the one-to-one corresponding relationship have the same pixel coordinates in the image.

[0057] Further, if the data sources of the CT image and the SPECT image are different, the resolutions of the CT image and the SPECT image can be different, and therefore, the non-rigid registration algorithm is needed to align the coordinate system of the CT image and the SPECT image without affecting the image quality of the CT image and the SPECT image, so that a one-to-many, many-to-one or many-to-many corresponding relationship exists between the pixel points of the CT image and the pixel points of the SPECT image.

[0058] S11, extracting the functional distribution feature of the SPECT image and the structural feature of the CT image through a preset deep learning model, and determining the artifact probability value of each pixel point in the SPECT image according to the functional distribution feature and the structural feature;

[0059] S12, dividing the pixel points in the SPECT image into an artifact category and a non-artifact category by comparing the artifact probability threshold value and the artifact probability value of each pixel point, and generating a binary mask image of the SPECT image according to the classification result of the pixel points.

[0060] The preset deep learning model can be a model dedicated to identifying the correlation between the CT image and the SPECT image. The functional distribution feature of the SPECT image can refer to a value or vector representing the functional distribution extracted by the preset deep learning model from the image region of the SPECT image. The structural feature of the CT image can refer to a value or vector representing the anatomical structure extracted by the preset deep learning model from the image region of the CT image. The artifact probability value of the pixel point can refer to the confidence or probability of the pixel point belonging to the artifact region predicted by the preset deep learning model based on the functional distribution feature and the structural feature.

[0061] The network of the preset deep learning model can be a convolutional neural network, a generative adversarial network, or a deep autoencoder. Specifically, when the network of the preset deep learning model is a convolutional neural network, the CT image and the SPECT image can be taken as inputs to identify the artifact region in the SPECT image and correct the artifact according to the structural information of the CT image (adjust the region in the SPECT image that does not conform to the structural information in the output layer of the network to achieve artifact elimination). In the training process of the convolutional neural network, the CT image can be used as prior knowledge to supervise the convolutional neural network to learn artifact identification and correction. When the network of the preset deep learning model is a generative adversarial network, the generator network accepts the CT image and the SPECT image as inputs to generate a SPECT image without artifacts, and the discriminator is used to determine whether the artifact-free SPECT image generated by the generator is a real SPECT image. The generator and the discriminator are in mutual opposition to generate more accurate and natural SPECT images. When the network of the preset deep learning model is a deep autoencoder, the CT image and the SPECT image are encoded into latent space representation by the encoder to extract high-dimensional features of the CT image and the SPECT image, respectively, and the corrected SPECT image is generated by the decoder. Compared with the optimization algorithm for artifact correction of the SPECT image,

[0062] The specific network of the preset deep learning model used in this embodiment is an encoder-decoder structure network, such as U-Net (a deep learning architecture) or its variants (such as Swin-UNet (a medical image segmentation model based on Transformer)). This encoder-decoder structure can effectively capture multi-scale features of the CT image and the SPECT image and perform accurate pixel-level positioning and segmentation.

[0063] The training data of the preset deep learning model can include: a plurality of groups of SPECT images and CT images paired with the SPECT images, each group of SPECT images being labeled with a plurality of functional regions, and each group of CT images being labeled with a plurality of structural regions. In the training process of the preset deep learning model, the deep learning model can learn the correct feature association relationship between the structural regions of the CT images and the functional regions of the SPECT images through a large amount of training data, so that the model can infer and correct the artifact region according to the structural information of the CT images, and the training optimization of the deep learning model is performed through the cross-entropy loss function. When the deep learning model is trained, the preset deep learning model of the embodiment is obtained, and in the actual application of the preset deep learning model, if the association relationship between a certain structural region of the CT image and the corresponding functional region of the SPECT image identified by the preset deep learning model is different from the correct association relationship learned during training, it can be considered that there is an artifact in the functional region of the SPECT image.

[0064] The embodiment introduces a deep learning model and combines the structural information of the CT image to realize efficient identification and accurate correction of artifacts in the SPECT image. Moreover, the deep neural network used in the embodiment can automatically learn the mapping relationship between structure and function through a large amount of SPECT / CT paired data, eliminating the dependence on prior assumptions, improving the generalization ability and robustness of the preset deep learning model, and at the same time, the CT image is embedded into the correction process of the SPECT image as structural guidance information, effectively improving the structural consistency and functional accuracy of the image, and can be quickly output through forward reasoning, meeting the real-time processing needs of the clinic.

[0065] Specifically, the trained preset deep learning model can have a dual-channel input, one channel being an input channel of the CT image and the other channel being an input channel of the SPECT image. The spatial dimension of the image feature map is gradually reduced through the encoder (a plurality of down-sampling layers) to extract the functional distribution features of the CT image and the structural features of the CT image, and the functional distribution features and the structural features are fused to obtain fused high-level features. The preset deep learning model identifies the association relationship between the CT image and the SPECT image from the fused high-level features, and the fused high-level features are up-sampled and refined through the decoder (a plurality of up-sampling layers), so that the preset deep learning model obtains more rich feature representation of the image for subsequent pixel point classification.

[0066] At the output layer of the preset deep learning model, an output is outputted as a probability map with the same size as the SPECT image (pixel-to-pixel correspondence), and the pixel value of each pixel in the probability map can represent the artifact probability value (0-1) of each pixel in the SPECT image. The probability map is converted into a mask map by setting an artifact probability threshold, so as to realize the identification of the artifact region, wherein the artifact probability threshold can be flexibly set according to actual conditions.

[0067] Specifically, in the embodiment, each pixel in the SPECT image is divided into an artifact category and a non-artifact category by the artifact probability threshold, and a mask map of the SPECT image is generated according to the classification result of the pixel. For each pixel in the SPECT image, if the artifact probability value of the pixel is greater than the artifact probability threshold, the category of the pixel is the artifact category, and if the artifact probability value of the pixel is not greater than the artifact probability threshold, the category of the pixel is the non-artifact category, so as to obtain the classification result of the pixel in the SPECT image. Each pixel in the SPECT image is traversed, if the category of the pixel is the artifact category, the pixel value of the mask pixel corresponding to the pixel is set to 1 (the image color is white), if the category of the pixel is the non-artifact category, the pixel value of the mask pixel corresponding to the pixel is set to 0 (the image color is black), so as to obtain a binary mask map.

[0068] S13, at least one artifact region in the SPECT image is determined by the binary mask map, and in the artifact region, the pixel value of the artifact region is corrected by the structural information of the CT image;

[0069] S14, the corrected artifact region and the non-artifact region of the SPECT image are fused to obtain a corrected SPECT image.

[0070] The artifact region can refer to an image region containing artifacts determined from the SPECT image according to the binary mask map. In the embodiment, the artifact region is determined from the SPECT image according to the information in the binary mask map. Specifically, since the binary mask map and the SPECT image have the same size (pixel-to-pixel correspondence), and in the binary mask map, the pixel value of the mask pixel in the artifact region is set to 1, and the pixel value of the mask pixel in the non-artifact region is set to 0, therefore, in the embodiment, the corresponding pixel in the SPECT image can be determined according to the position of the mask pixel with the pixel value of 1 in the binary mask map, so as to determine the artifact region in the SPECT image composed of these pixels (the pixel value of the corresponding mask pixel is 1). In artifact correction, the binary mask map can directly guide the repair network to correct only the artifact region composed of these pixels (the pixel value of the corresponding mask pixel is 1) in the SPECT image.

[0071] Of course, in another alternative embodiment, the present embodiment can directly fuse the binary mask image with the SPECT image to obtain the artifact region in the SPECT image. Specifically, since the binary mask image and the SPECT image have the same size (pixel-to-pixel correspondence), the pixel value of the mask pixel of the binary mask image can be multiplied by the pixel value of the corresponding pixel in the SPECT image to extract the artifact region from the SPECT image. Therefore, when the binary mask image is multiplied by the pixel value of the SPECT image, the pixel value of the artifact region remains unchanged (multiplied by 1) in the SPECT image, and the pixel value of the non-artifact region is all 0 (multiplied by 0), thereby segmenting the artifact region in the SPECT image.

[0072] The present embodiment can correct the artifact region by using a trained repair network (such as a U-Net or a generative adversarial network). During training, the repair network also learns the correct association between the structural region of the CT image and the functional region of the SPECT image from a large amount of training data, so as to correct the artifact according to the structural information of the CT image. In addition, the repair network also needs to train the image correction quality, score the image output by the repair network each time during training, or calculate the difference between the corrected image and the reference true value image using L1 / L2 loss, so as to feed back the optimization or adjust the parameters of the repair network, improve the image correction quality, and also introduce a structural similarity loss to improve the image quality and structural preservation. When the repair network is a generative adversarial network, the discriminator in the generative adversarial network can be used to determine whether the corrected region is natural and consistent with the surrounding image structure.

[0073] Specifically, the input of the trained repair network can include: an original SPECT image, a CT image, and a binary mask image, the binary mask image guides the repair network to correct only the artifact region of the original SPECT image, and the repair network corrects the pixel value of the artifact region by using the structural information of the CT image and the image association relationship (the feature association relationship between the functional region in the SPECT image and the structural region in the CT image learned during training). Since the SPECT image is a grayscale image, the present embodiment selects to adjust the grayscale value of the pixel in the artifact region to correct the artifact region of the SPECT image.

[0074] After the artifact region is corrected, the repair network fuses the corrected artifact region with the non-artifact region of the original SPECT image at the pixel level to realize the region splicing of the corrected artifact region and the non-artifact region of the original SPECT image, and obtains a complete SPECT image with artifact correction. Of course, in another alternative embodiment, the CT image and the SPECT image with artifact correction can be fused to obtain a SPECT / CT fusion image and display it, which is convenient for doctors to further judge the consistency of the structure and function of the tissue or organ.

[0075] Further, after obtaining the artifact-corrected SPECT image, the embodiment can evaluate the image quality of the artifact-corrected SPECT image. If the image quality does not meet the requirements, the artifact-corrected SPECT image is input again to the repair network for secondary iteration correction. The embodiment can select a full-reference image quality evaluation index to evaluate the image quality of the artifact-corrected SPECT image, such as PSNR (peak signal-to-noise ratio) and SSIM (structural similarity index). Specifically, the embodiment can select a reference “gold standard” SPECT image (which can be a high-quality SPECT image in the training data), calculate the PSNR value and SSIM value between the “gold standard” SPECT image and the artifact-corrected SPECT image, and if the PSNR value and SSIM value do not meet the threshold value, it can be considered that the image quality of the artifact-corrected SPECT image does not meet the requirements. For example, the PSNR threshold is set to 30 dB (decibels), and the SSIM threshold is set to 0.95. When the PSNR value is greater than 30 dB and the SSIM value is greater than 0.95, the image quality is excellent, otherwise the image quality is not excellent. Of course, in another alternative implementation, a no-reference image quality evaluation index (such as NIQE, BRISQUE) or a discriminator network trained to evaluate the image quality of the artifact-corrected SPECT image can also be used. NIQE (Natural Image Quality Evaluator) is a no-reference image quality evaluation algorithm that analyzes the statistical characteristics of distorted images to predict the subjective quality score of the image. BRISQUE (Blind / Referenceless Image Spatial Quality Evaluator) is a no-reference image quality evaluation algorithm that analyzes the statistical regularity deviation of natural images to quantify image quality.

[0076] This embodiment provides an image correction method for SPECT images. The method acquires SPECT and CT images of the same patient and aligns them to establish a correspondence between pixels in the SPECT and CT images. It extracts the functional distribution features of the SPECT image and the structural features of the CT image, and determines the artifact probability value of each pixel in the SPECT image based on these features. Each pixel in the SPECT image is then categorized into artifact and non-artifact categories using an artifact probability threshold, and a binary mask image of the SPECT image is generated based on the classification results. At least one artifact region in the SPECT image is identified using the binary mask image, and the pixel values ​​in the artifact region are corrected using the structural information from the CT image. The corrected artifact region is then fused with the non-artifact region of the SPECT image to obtain the corrected SPECT image. This method learns and utilizes the image information correlation between the CT and SPECT images to assist in identifying and correcting artifact regions in the SPECT image. By utilizing the actual structural information of the CT image to correct the SPECT image, it eliminates dependence on prior assumptions and effectively improves the accuracy of SPECT image correction. This embodiment also integrates quality assessment and feedback mechanisms, constructs a closed-loop optimization process, and further enhances the stability and intelligence of SPECT image correction effects.

[0077] The above describes an image correction method for SPECT images provided by embodiments of this application. The following will describe a system that applies the above-described image correction method for SPECT images.

[0078] Please see Figure 2 , Figure 2 This is a schematic diagram of the structure of an image correction system for SPECT images provided in an embodiment of this application. Figure 2 As shown, the image correction system for the SPECT image may include:

[0079] The acquisition unit 100 is used to acquire single-photon emission computed tomography (SPECT) images and computed tomography (CT) images of the same patient, and align the coordinate systems of the SPECT images and the CT images so that the pixels of the SPECT images and the pixels of the CT images have a corresponding relationship.

[0080] Analysis unit 110 is used to extract the functional distribution features of SPECT images and the structural features of CT images through a preset deep learning model, and to determine the artifact probability value of each pixel in the SPECT image based on the functional distribution features and structural features.

[0081] The classification unit 120 is configured to divide the pixel points in the SPECT image into a pseudo artifact category and a non-pseudo artifact category by comparing the pseudo artifact probability threshold value with the pseudo artifact probability value of each pixel point, and generate a binary mask image of the SPECT image according to the classification result of the pixel points.

[0082] The correction unit 130 is configured to determine at least one pseudo artifact region in the SPECT image through the binary mask image, and correct the pixel value of the pseudo artifact region through the structural information of the CT image in the pseudo artifact region.

[0083] The fusion unit 140 is configured to fuse the corrected pseudo artifact region with the non-pseudo artifact region of the SPECT image to obtain a corrected SPECT image.

[0084] In a possible implementation, the acquisition unit 100 is configured to acquire the single photon emission computed tomography (SPECT) image and the computed tomography (CT) image of the same patient, and can be specifically configured as follows:

[0085] The SPECT image and the CT image of the same patient are simultaneously acquired through a SPECT / CT integrated machine.

[0086] In a possible implementation, the training data of the preset deep learning model can include a plurality of groups of SPECT images and CT images paired with the SPECT images, and each group of SPECT images is labeled with a plurality of functional regions, and each group of CT images is labeled with a plurality of structural regions.

[0087] In a possible implementation, the classification unit 120 is configured to divide the pixel points in the SPECT image into a pseudo artifact category and a non-pseudo artifact category by comparing the pseudo artifact probability threshold value with the pseudo artifact probability value of each pixel point, and can be specifically configured as follows:

[0088] For each pixel point in the SPECT image, if the pseudo artifact probability value of the pixel point is greater than the pseudo artifact probability threshold value, the category of the pixel point is the pseudo artifact category, and if the pseudo artifact probability value of the pixel point is not greater than the pseudo artifact probability threshold value, the category of the pixel point is the non-pseudo artifact category; and the classification result of the pixel points in the SPECT image is obtained.

[0089] In a possible implementation, the mask pixel points of the binary mask image correspond to the pixel points of the SPECT image one by one.

[0090] The classification unit 120 is configured to generate a binary mask image of the SPECT image according to the classification result of the pixel points, and can be specifically configured as follows:

[0091] Iterate through each pixel in the SPECT image. If the pixel belongs to the artifact category, set the pixel value of the corresponding mask pixel to 1; if the pixel belongs to the non-artifact category, set the pixel value of the corresponding mask pixel to 0; thus obtaining a binary mask image.

[0092] In one possible implementation, the correction unit 130 corrects the pixel values ​​of the artifact region using the structural information of the CT image, which can be specifically configured as follows:

[0093] The grayscale values ​​of artifact regions are corrected by using structural information and image correlations in CT images. The image correlations are: the feature correlations between functional regions in SPECT images and structural regions in CT images.

[0094] In one possible implementation, the system may also include an evaluation unit:

[0095] The evaluation unit is used to evaluate the image quality of the corrected SPECT image. If the image quality does not meet the requirements, the corrected SPECT image is then subjected to a second iteration for correction.

[0096] In one possible implementation, the system also includes constraint units:

[0097] Constraint units are used to use the structural information of CT images as constraints in the reconstruction process of SPECT images.

[0098] This application also provides an electronic device in its embodiments. (See reference...) Figure 3 The diagram illustrates a structural schematic suitable for implementing the electronic device in the embodiments of this application. The electronic device in the embodiments of this application may include, but is not limited to, fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 3 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0099] like Figure 3 As shown, the electronic device may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. When the electronic device is powered on, the RAM 303 also stores various programs and data required for the operation of the electronic device. The processing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0100] In general, the following devices can be connected to the I / O interface 305: input devices 306 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 308 including, for example, a memory card, a hard disk, etc.; and communication devices 309. The communication devices 309 can allow the electronic device to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 The electronic device is shown with various devices, but it is understood that all of the shown devices are not required to be implemented or present. More or fewer devices can alternatively be implemented or present.

[0101] The embodiment of the present application further provides a computer program product comprising computer readable instructions, which, when executed on an electronic device, cause the electronic device to implement any of the image correction methods of SPECT images provided by the embodiments of the present application.

[0102] The embodiment of the present application further provides a computer readable storage medium, which carries one or more computer programs, and when the one or more computer programs are executed by an electronic device, the electronic device can implement any of the image correction methods of SPECT images provided by the embodiments of the present application.

[0103] In addition, it should be noted that the system embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e. they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments. In addition, the connection relationship between the modules in the system embodiments provided by the present application indicates that there is a communication connection between them, which can be implemented as one or more communication buses or signal lines.

[0104] Those skilled in the art can clearly understand that the application can be implemented by means of software plus necessary universal hardware, and of course can also be implemented by means of dedicated hardware including special-purpose integrated circuits, special-purpose CPUs, special-purpose memories, special-purpose components, etc. Generally, any function completed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure for implementing the same function can also be various, such as analog circuits, digital circuits, or special-purpose circuits, etc. However, for the present application, software program implementation is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a readable storage medium, such as a floppy disk, a U disk, a mobile hard disk, a ROM, a RAM, a magnetic disk or an optical disk, etc., and includes a plurality of instructions for causing a computer device (which can be a personal computer, a training device, or a network device, etc.) to execute the methods described in various embodiments of the present application.

[0105] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments can be implemented in the form of a computer program product.

[0106] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another, for example, the computer instructions can be transmitted from one website, computer, training device or data center to another website, computer, training device or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that can be stored by a computer or a data storage device such as a training device, a data center, etc. integrated with one or more available media. The available media can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid state disk (SSD)), etc.

[0107] Each of the embodiments in the specification is described in a relevant manner, and the same or similar parts between the embodiments can be referred to each other. Each of the embodiments focuses on the difference from other embodiments. In particular, for the device embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiments.

[0108] It can be understood that, before using the technical solutions disclosed in the embodiments of the present disclosure, the type of personal information involved in the present disclosure, the use range, the use scenario, etc. should be informed to the user and the authorization of the user should be obtained through appropriate means according to relevant laws and regulations.

[0109] The above only describes the embodiments of the present application and is not intended to limit the present application. The present application can have various changes and modifications for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for image correction of SPECT images, characterized in that, The image correction method for the SPECT image includes: Acquire single-photon emission computed tomography (SPECT) images and computed tomography (CT) images of the same patient, and align the coordinate systems of the SPECT images and the CT images so that the pixels of the SPECT images and the pixels of the CT images have a corresponding relationship. The functional distribution features of the SPECT image and the structural features of the CT image are extracted by a preset deep learning model, and the artifact probability value of each pixel in the SPECT image is determined based on the functional distribution features and the structural features. By comparing the artifact probability threshold and the artifact probability value of each pixel, the pixels in the SPECT image are divided into artifact category and non-artifact category, and a binary mask image of the SPECT image is generated based on the classification results of the pixels. At least one artifact region in the SPECT image is determined by the binary mask image, and the pixel values ​​of the artifact region are corrected by the structural information of the CT image. The corrected artifact region is fused with the non-artifact region of the SPECT image to obtain the corrected SPECT image.

2. The image correction method for SPECT images according to claim 1, characterized in that, The acquisition of single-photon emission computed tomography (SPECT) images and computed tomography (CT) images of the same patient includes: The SPECT and CT images of the same patient can be acquired simultaneously using a SPECT / CT integrated machine.

3. The image correction method for SPECT images according to claim 1, characterized in that, The training data for the preset deep learning model includes: multiple sets of SPECT images and CT images paired with the SPECT images. In each set, the SPECT images are labeled with multiple functional regions, and in each set, the CT images are labeled with multiple structural regions.

4. The image correction method for SPECT images according to claim 1, characterized in that, The process of classifying pixels in the SPECT image into artifact and non-artifact categories by comparing an artifact probability threshold with the artifact probability value of each pixel includes: For each pixel in the SPECT image, if the artifact probability value of the pixel is greater than the artifact probability threshold, then the pixel is classified as the artifact category; if the artifact probability value of the pixel is not greater than the artifact probability threshold, then the pixel is classified as the non-artifact category. Obtain the classification results of the pixels in the SPECT image.

5. The image correction method for SPECT images according to claim 1, characterized in that, The mask pixels of the binary mask image correspond one-to-one with the pixels of the SPECT image; The step of generating a binary mask image of the SPECT image based on the pixel classification results includes: Traverse each pixel in the SPECT image. If the pixel belongs to the artifact category, set the pixel value of the mask pixel corresponding to the pixel to 1. If the category of the pixel is the non-artifact category, then the pixel value of the mask pixel corresponding to the pixel is set to 0; Obtain the binary mask image.

6. The image correction method for SPECT images according to claim 1, characterized in that, The step of correcting the pixel values ​​of the artifact region using the structural information of the CT image includes: The grayscale value of the artifact region is corrected by using the structural information and image correlation of the CT image. The image correlation is the feature correlation between the functional regions in the SPECT image and the structural regions in the CT image.

7. The image correction method for SPECT images according to claim 1, characterized in that, Also includes: The image quality of the corrected SPECT image is evaluated. If the image quality does not meet the requirements, a second iteration of correction is performed on the corrected SPECT image.

8. The image correction method for SPECT images according to claim 1, characterized in that, Also includes: During the reconstruction of the SPECT image, the structural information of the CT image is used as a constraint condition for the reconstruction process.

9. An image correction system for SPECT images, characterized in that, The image correction system for the SPECT image includes: The acquisition unit is used to acquire single-photon emission computed tomography (SPECT) images and computed tomography (CT) images of the same patient, and align the coordinate systems of the SPECT images and the CT images so that the pixels of the SPECT images and the pixels of the CT images have a corresponding relationship. The analysis unit is used to extract the functional distribution features of the SPECT image and the structural features of the CT image through a preset deep learning model, and to determine the artifact probability value of each pixel in the SPECT image based on the functional distribution features and the structural features. The classification unit is used to classify the pixels in the SPECT image into artifact categories and non-artifact categories by comparing the artifact probability threshold and the artifact probability value of each pixel, and to generate a binary mask image of the SPECT image based on the classification results of the pixels. The correction unit is used to determine at least one artifact region in the SPECT image through the binary mask image, and to correct the pixel value of the artifact region through the structural information of the CT image. The fusion unit is used to fuse the corrected artifact region with the non-artifact region of the SPECT image to obtain the corrected SPECT image.

10. An electronic device, characterized in that, It includes at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is configured to execute the computer program to enable the electronic device to implement the image correction method for SPECT images as described in any one of claims 1 to 8.

Images