Thyroid section image quality evaluation method and device, equipment and storage medium

Abstract

The embodiment of the invention provides a thyroid section image quality evaluation method and device, equipment and a storage medium, and belongs to the technical field of image processing. The method comprises the following steps: acquiring a cell slice image; wherein the cell section image is obtained by scanning thyroid cell sections by a scanner; performing cell image extraction on the cell slice image to obtain a selected cell image; inputting the selected cell image into a preset thyroid cell information detection model for image content detection to obtain cell image content information; wherein the cell image content information comprises cell cluster information and image impurity information; and inputting the cell mass information and the image impurity information into a preset slice quality evaluation model to perform quality evaluation to obtain a thyroid cell quality evaluation result of the cell slice image. According to the embodiment of the invention, the quality of the scanned slice image can be accurately evaluated, and a high-quality thyroid cell slice image is obtained.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and particularly to a method and device, equipment, and storage medium for evaluating the quality of thyroid section images. Background Art

[0002] Thyroid section images are obtained by scanning thyroid sections using a scanner, counting the number of cells in the scanned images, and finally evaluating the quality of the thyroid sections based on the cell count results. However, due to the large amount of impurities in thyroid sections, it affects the accuracy of the quality evaluation of the scanned section images. Therefore, how to accurately evaluate the quality of scanned section images has become an urgent problem to be solved. Summary of the Invention

[0003] The main purpose of the embodiments of this application is to propose a method and device, equipment, and storage medium for evaluating the quality of thyroid section images, aiming to accurately evaluate the quality of scanned section images and obtain high-quality thyroid cell section images.

[0004] To achieve the above object, in the first aspect of the embodiments of this application, a method for evaluating the quality of thyroid section images is proposed. The method includes: Obtain a cell section image; wherein, the cell section image is obtained by scanning a thyroid cell section using a scanner; Extract cell images from the cell section image to obtain selected cell images; Input the selected cell images into a preset thyroid cell information detection model for image content detection to obtain cell image content information; wherein, the cell image content information includes: cell cluster information and image impurity information; Input the cell cluster information and the image impurity information into a preset section quality evaluation model for quality evaluation to obtain the thyroid cell quality evaluation result of the cell section image.

[0005] In some embodiments, the cell cluster information includes thyroid follicular epithelial cell cluster information and impurity cell cluster information; the inputting the cell cluster information and the image impurity information into a preset section quality evaluation model for quality evaluation to obtain the thyroid cell quality evaluation result of the cell section image includes: Obtain the confidence of the thyroid follicular epithelial cell cluster information to obtain thyroid follicular epithelial cell cluster confidence information; Screen the thyroid follicular epithelial cell cluster information according to a preset confidence threshold and the thyroid follicular epithelial cell cluster confidence information to obtain selected cell cluster information; Input the selected cell cluster information, the impurity cell cluster information, and the image impurity information into the slice quality assessment model for cell cluster quantity statistics to obtain the quantity of thyroid cell clusters. Determine the thyroid cell quality assessment result of the cell slice image according to the preset thyroid cell cluster determination rule and the quantity of thyroid cell clusters.

[0006] In some embodiments, the extracting cell images from the cell slice image to obtain selected cell images includes: Clean the blood area in the cell slice image through a preset image blood area cleaning model to obtain a cell blood cleaning image; Perform binary segmentation on the cell blood cleaning image to obtain a binary cell image; Perform image cropping on the binary cell image to obtain selected cell images.

[0007] In some embodiments, the performing binary segmentation on the cell blood cleaning image to obtain a binary cell image includes: Obtain the number of pixel points of the cell blood cleaning image, and obtain the pixel gray value of each pixel point in the cell blood cleaning image; Classify the multiple pixel gray values according to a preset multiple gray value thresholds to obtain a pixel classification group pair corresponding to each gray value threshold; wherein, the pixel classification group pair includes a first pixel group and a second pixel group; Obtain the ratio of each first pixel group to the number of pixel points to obtain a first ratio, and obtain the ratio of each second pixel group to the number of pixel points to obtain a second ratio; Perform aggregation calculation on the second pixel group corresponding to each first pixel group, the first ratio, and the second ratio to obtain the pixel variance value of each gray value threshold; Obtain the maximum pixel variance value to obtain a segmentation variance value, and obtain the corresponding gray value threshold according to the segmentation variance value to obtain a segmentation threshold; Perform image binary processing on the cell blood cleaning image according to the segmentation threshold to obtain the binary cell image.

[0008] In some embodiments, the image blood area cleaning model includes a color cleaning sub-model and a morphological cleaning sub-model; the cleaning the blood area in the cell slice image through a preset image blood area cleaning model to obtain a cell blood cleaning image includes: Perform color screening on the cell slice image through the color cleaning sub-model to obtain a selected color image; wherein, the selected color image represents an image without blood color; The selected color image is subjected to morphological screening by the morphological cleaning sub-model to obtain the cell blood cleaning image; wherein, the cell blood cleaning image represents an image without blood morphology.

[0009] In some embodiments, the binarized cell image is subjected to image cropping to obtain a selected cell image, including: The binarized cell image is scaled according to a preset image scaling ratio to obtain a selected scaled image; The selected scaled image is cropped according to a preset cropping window to obtain the selected cell image.

[0010] In some embodiments, after the cell mass information and the image impurity information are input into a preset slice quality evaluation model for quality evaluation to obtain the thyroid cell quality evaluation result of the cell slice image, it further includes: If the thyroid cell quality evaluation result indicates that the number of thyroid cell masses conforms to the thyroid cell mass determination rule, slice qualified information is output; If the thyroid cell quality evaluation result indicates that the number of thyroid cell masses does not conform to the thyroid cell mass determination rule, the selected scaled image is subjected to image sliding cropping according to a preset sliding window step size and the cropping window to obtain an updated image; the updated image is input into the thyroid cell information detection model for image content detection to obtain updated image content information; wherein, the updated image content information includes cell mass update information and impurity update information; the cell mass update information and the impurity update information are input into the slice quality evaluation model to perform quality evaluation on the updated image to obtain an updated cell quality evaluation result.

[0011] To achieve the above object, a second aspect of the embodiments of the present application proposes a thyroid slice image quality evaluation device, the device includes: An acquisition module, configured to acquire a cell slice image; wherein, the cell slice image is obtained by a scanner scanning a thyroid cell slice; An extraction module, configured to extract a cell image from the cell slice image to obtain a selected cell image; A detection module, configured to input the selected cell image into a preset thyroid cell information detection model for image content detection to obtain cell image content information; wherein, the cell image content information includes: cell mass information and image impurity information; An evaluation module, configured to input the cell mass information and the image impurity information into a preset slice quality evaluation model for quality evaluation to obtain the thyroid cell quality evaluation result of the cell slice image.

[0012] To achieve the above object, a third aspect of the embodiments of the present application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the method described in the first aspect above is implemented.

[0013] To achieve the above object, a fourth aspect of the embodiments of the present application provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect above is implemented.

[0014] A method, device, equipment and storage medium for evaluating the quality of thyroid slice images proposed in the present application extract cell images from the cell slice images obtained by scanning thyroid cell slices through a scanner, extract the image content related to thyroid cells in the cell slice images to obtain selected cell images, so as to exclude the interference information in the images and avoid the interference of the interference information in the images on subsequent image processing. Then, the selected cell images are input into a thyroid cell information detection model for image content detection to obtain cell cluster information and image impurity information. Subsequently, the cell cluster information and image impurity information are input into a slice quality evaluation model for quality evaluation to obtain the quality evaluation result of thyroid cells. By inputting the image impurity information and cell cluster information into the slice quality evaluation model simultaneously and combining the cell impurity information and cell cluster information to evaluate the quality of thyroid cell slices, the amount of information input to the slice quality model is increased, thereby improving the accuracy of the quality evaluation of cell slice images by the slice quality evaluation model. That is, the method for evaluating the quality of thyroid slice images provided by the present disclosure can improve the accuracy of the quality evaluation of scanned slice images and obtain high-quality thyroid cell slice images. Description of the Drawings

[0015] Figure 1 is a flowchart of the method for evaluating the quality of thyroid slice images provided by the embodiments of the present application; Figure 2 is Figure 1 a flowchart of step S102 in Figure 3 is Figure 2 a flowchart of step S201 in Figure 4 is Figure 2 a flowchart of step S202 in Figure 5 is Figure 2 a flowchart of step S203 in Figure 6 is Figure 1 a flowchart of step S104 in Figure 7It is a flowchart of a method for evaluating the quality of thyroid section images provided by another embodiment of the present application; Figure 8 It is a schematic structural diagram of a device for evaluating the quality of thyroid section images provided by an embodiment of the present application; Figure 9 It is a schematic hardware structure diagram of an electronic device provided by an embodiment of the present application; Figure 10 It is a schematic diagram of image cropping provided by an embodiment of the present application; Figure 11 It is an overall flowchart of a method for evaluating the quality of thyroid section images provided by an embodiment of the present application. Detailed implementation manners

[0016] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0017] It should be noted that although functional module division is performed in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order from the module division in the device or the order in the flowchart. Terms such as "first" and "second" in the specification, claims and the above-mentioned drawings are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0019] First, several nouns involved in the present application are analyzed: Whole Slide Image: It is a high-resolution two-dimensional image of a cell sample obtained by a scanner, a biomedical image used to study cell structure and function. The whole slide image belongs to the interdisciplinary field of biomedical imaging and cell biology. It is an image obtained by scanning a cell sample through a scanner to study cell structure and function, used to observe and analyze cell morphology, tissue structure, pathological changes, and assist in scientific research and medical diagnosis. This technology has important applications in the fields of biomedical research, clinical pathology, cancer research, and drug development.

[0020] Degenerate Cell Mass: It is a mass of cells that have lost their normal function and structure, usually due to disease or aging. Degenerate cell masses belong to the research fields of biomedicine and pathology, mainly used to study the mechanisms of cell degeneration and pathological processes, help diagnose and treat various diseases, and have important applications in the fields of clinical pathology, cancer research, neurodegenerative disease research, and regenerative medicine.

[0021] Neutrophil Cluster: It is a cluster of neutrophils formed by the aggregation of neutrophils in the body, usually in response to infection or inflammation. Neutrophil clusters belong to the research fields of immunology and pathology, mainly used to study and understand immune responses and their roles in anti-infection and inflammatory processes. Neutrophil clusters have important applications in the fields of clinical diagnosis, infectious disease research, inflammatory disease research, and immunotherapy.

[0022] Interstitial Fluid: It is the liquid component in cell tissues, belonging to the categories of biology and medicine, containing water, electrolytes, nutrients, and metabolites, etc., and is an important medium for the exchange of substances inside and outside cells. In biology, the study of interstitial fluid helps to understand the effects of the internal and external environments of cells on the functions and metabolism of organisms, as well as the interaction mechanisms between cells. In the medical field, the analysis of interstitial fluid can be used to diagnose diseases, monitor physiological states, and evaluate treatment effects. For example, blood, lymph fluid, and interstitial fluid between cells are all important interstitial fluids.

[0023] Confidence Level: In statistics, it refers to the degree of estimation of the probability of a certain event occurring. In the fields of data analysis and machine learning, the confidence level is used to measure the confidence of the model in the prediction results. A high confidence level indicates that the model has a high degree of certainty in the prediction results, while a low confidence level indicates that the model has a high degree of uncertainty in the results. In practical applications, the concept of confidence level is widely used in decision-making, risk assessment, and the reliability analysis of prediction results, etc.

[0024] Thyroid slice images are obtained by scanning thyroid slices using a scanner, counting the number of cells in the scanned images, and finally evaluating the quality of the thyroid slices based on the cell count results. However, due to the large amount of impurities in the thyroid slices, it affects the accuracy of the quality assessment of the scanned slice images. Therefore, how to accurately evaluate the quality of the scanned slice images has become an urgent problem to be solved.

[0025] Based on this, the embodiments of this application provide a method and device, equipment, and storage medium for evaluating the quality of thyroid slice images, aiming to accurately evaluate the quality of the scanned slice images and obtain high-quality thyroid cell slice images.

[0026] A method, device, equipment and storage medium for evaluating the quality of thyroid slice images provided by an embodiment of the present application will be specifically described through the following embodiments. First, the method for evaluating the quality of thyroid slice images in the embodiments of the present application will be described.

[0027] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0028] The basic technologies of artificial intelligence generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. The software technologies of artificial intelligence mainly include several major directions such as computer vision technology, robotics technology, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0029] The method for evaluating the quality of thyroid slice images provided by the embodiments of the present application relates to the field of image processing technology. The method for evaluating the quality of thyroid slice images provided by the embodiments of the present application can be applied to a terminal, or to a server side, or can also be software running on a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, or can also be configured as a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the method for evaluating the quality of thyroid slice images, etc., but is not limited to the above forms.

[0030] This application can be used in numerous general-purpose or special-purpose computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0031] Figure 1 FIG. is an alternative flowchart of the thyroid slice image quality assessment method provided by an embodiment of this application. Figure 1 The method in may include but is not limited to steps S101 to S104.

[0032] Step S101, obtain a cell slice image; wherein, the cell slice image is obtained by a scanner scanning a thyroid cell slice. Step S102, extract cell images from the cell slice image to obtain selected cell images. Step S103, input the selected cell images into a preset thyroid cell information detection model for image content detection to obtain cell image content information; wherein, the cell image content information includes: cell cluster information and image impurity information. Step S104, input the cell cluster information and the image impurity information into a preset slice quality assessment model for quality assessment to obtain the thyroid cell quality assessment result of the cell slice image.

[0033] Steps S101 to S104 illustrated in the embodiments of the present application perform cell image extraction on the cell section image obtained by scanning a thyroid cell section with a scanner, extract the image content related to thyroid cells in the cell section image to obtain a selected cell image, thereby excluding the interference information in the image, avoiding interference of the interference information in the image on subsequent image processing, and inputting the selected cell image into a thyroid cell information detection model for image content detection to obtain cell mass information and image impurity information. Then, the cell mass information and image impurity information are input into a section quality evaluation model for quality evaluation, and the quality evaluation result of thyroid cells is obtained. The image impurity information and cell mass information are input into the section quality evaluation model at the same time, and the thyroid cell section is evaluated for quality by combining the cell impurity information and cell mass information, increasing the amount of information input to the section quality model, thereby improving the accuracy of the quality evaluation of the cell section image by the section quality evaluation model. That is, the thyroid section image quality evaluation method provided by the present disclosure can improve the accuracy of the quality evaluation of the scanned section image and obtain a high-quality thyroid cell section image.

[0034] In step S101 of some embodiments, a thyroid tissue is punctured and made into a section to obtain a thyroid section, and the thyroid section is scanned with a scanner to obtain a thyroid cell section image, that is, a cell section image.

[0035] Please refer to Figure 2 , in some embodiments, step S102 may include but is not limited to steps S201 to S203: Step S201, cleaning the blood area in the cell section image through a preset image blood area cleaning model to obtain a cell blood cleaning image; Step S202, performing binary segmentation on the cell blood cleaning image to obtain a binary cell image; Step S203, performing image cropping on the binary cell image to obtain a selected cell image.

[0036] Steps S201 to S203 illustrated in the embodiments of the present application perform blood area removal processing on the cell section image to obtain a cell blood cleaning image, avoiding interference of the blood area on subsequent quality evaluation, performing binary segmentation on the cell blood cleaning image to obtain a binary cell image, adaptively determining the segmentation threshold, accurately extracting and segmenting thyroid cells, removing the irrelevant image content in the image, reducing the area that needs to be processed for image processing, improving the speed of quality evaluation of the thyroid section image, reducing the area that needs to be processed for image processing, and finally performing image cropping on the binary cell image to obtain a selected cell image to obtain the effective area of the binary cell image, accelerating the speed of quality evaluation of the cell section image.

[0037] Please refer to Figure 3 , in some embodiments, the image blood region cleaning model includes: a color cleaning sub-model and a morphological cleaning sub-model; step S201 may include but is not limited to steps S301 to S302: Step S301, perform color screening on the cell section image through the color cleaning sub-model to obtain a selected color image; wherein, the selected color image represents an image without blood color; Step S302, perform morphological screening on the selected color image through the morphological cleaning sub-model to obtain a cell blood cleaning image; wherein, the cell blood cleaning image represents an image without blood morphology.

[0038] Steps S301 to S302 shown in the embodiments of the present application remove the blood image in the cell section image through the color cleaning sub-model and the morphological cleaning sub-model, avoiding interference of the subsequent blood image on the quality assessment, thereby improving the accuracy of the quality assessment of the cell section image.

[0039] In step S301 of some embodiments, the image content in the cell section image that matches the blood color is removed through the color cleaning sub-model to obtain a cell blood cleaning image. Since during the production of thyroid sections, due to the influence of the puncture and section preparation processes, a small amount of blood will be contained in the thyroid sections, and the image obtained after passing through the scanner contains a blood region. Removing the blood region in the image can improve the quality of the cell section image and the accuracy of the subsequent quality assessment of the cell section image.

[0040] In one embodiment, the color cleaning sub-model is a Mask R-CNN model, which is used to extract the blood in the cell section image according to the blood color. Mask R-CNN is formed by sequentially connecting a preliminary prediction layer, a feature extraction layer, a fully connected classification layer, a bounding box regression layer, and a pixel prediction layer. The preliminary prediction layer is used to slide in the cell section image to generate multiple anchor boxes that may contain blood, and then the feature extraction layer extracts the image features of each anchor box. The fully connected classification layer determines whether the image in each anchor box is a blood image according to the image features to obtain a discrimination result. Furthermore, the bounding box regression layer corrects the boundary of each anchor box containing the blood image through the discrimination result, and then inputs the anchor box with the corrected boundary into the pixel prediction layer to judge each pixel point in the anchor box and output a binary matrix with the same size as the cell section image. The binary matrix represents whether each pixel point in the cell section image is a blood pixel point. According to the binary matrix, each blood pixel point in the cell section image is assigned a value of zero, completing the color screening operation of the cell section image according to the blood color.

[0041] In one embodiment, the color cleaning sub-model is a threshold segmentation model, which is used to perform threshold segmentation on the cell slice image according to a preset blood color threshold range. When the pixel point of the cell slice image is within the blood color threshold range, the pixel point is set to zero, thereby completing the color screening of the cell slice image according to the blood color.

[0042] In step S302 of some embodiments, the selected cell image is obtained by the morphological cleaning sub-model removing the image content in the cell slice image that does not conform to the morphology of thyroid cells but conforms to the morphology of blood.

[0043] In one embodiment, the morphological cleaning sub-model is an edge detection model, which detects and removes the image content in the image that does not conform to the morphology of thyroid cells but conforms to the morphology of blood to obtain a cell blood cleaning image.

[0044] In one embodiment, the morphological cleaning sub-model is a morphological opening model, which specifically includes performing an image erosion operation on the selected cell image to obtain an eroded image, and then performing an image dilation operation on the eroded image to obtain a cell blood cleaning image. In the selected cell image, blood occupies a small part of the area, and the morphological opening model can remove small objects in the image and smooth the object boundaries, thereby accurately removing the blood image.

[0045] Please refer to Figure 4 , in some embodiments, step S202 may include but is not limited to steps S401 to S406: Step S401, obtaining the number of pixel points of the cell blood cleaning image, and obtaining the pixel gray value of each pixel point in the cell blood cleaning image; Step S402, classifying the multiple pixel gray values according to a preset multiple gray value thresholds to obtain a pixel classification pair corresponding to each gray value threshold; wherein, the pixel classification pair includes a first pixel group and a second pixel group; Step S403, obtaining the ratio of each first pixel group to the number of pixel points to obtain a first ratio, and obtaining the ratio of each second pixel group to the number of pixel points to obtain a second ratio; Step S404, performing an aggregation calculation on the second pixel group, the first ratio, and the second ratio corresponding to each first pixel group to obtain a pixel variance value for each gray value threshold; Step S405, obtaining the maximum pixel variance value to obtain a segmentation variance value, and obtaining the corresponding gray value threshold according to the segmentation variance value to obtain a segmentation threshold; Step S406, performing image binarization processing on the cell blood cleaning image according to the segmentation threshold to obtain a binarized cell image.

[0046] In the steps S401 to S406 shown in the embodiments of the present application, variance processing is performed on the cell blood washing image according to a plurality of preset gray value thresholds, and the gray value threshold with the largest variance is obtained as the segmentation threshold, and the cell blood washing image is binarized according to the segmentation threshold, so as to realize the adaptive calculation of the pixel points of the image to obtain the optimal segmentation threshold, and improve the accuracy of image segmentation.

[0047] In step S401 of some embodiments, the cell blood washing image includes a plurality of pixel points. First, the number of pixel points of the cell blood washing image is obtained, and then the pixel gray value of each pixel point is obtained.

[0048] In step S402 of some embodiments, there are a total of 256 gray value thresholds, which are gray value thresholds from 0 to 255 respectively. The plurality of pixel gray values are classified according to each gray value threshold to obtain 256 groups of pixel classification teams.

[0049] In one embodiment, if the gray value threshold is 232, the plurality of pixel gray values are classified according to 232 to obtain a first pixel group and a second pixel group. Among them, the first pixel group is the set of pixel points in the cell blood washing image that are less than or equal to 232, and the second pixel group is the set of pixel points in the cell blood washing image that are greater than 232.

[0050] In one embodiment, if the gray value threshold is 200, the plurality of pixel gray values are classified according to 200 to obtain a first pixel group and a second pixel group. Among them, the first pixel group is the set of pixel points in the cell blood washing image that are less than or equal to 200, and the second pixel group is the set of pixel points in the cell blood washing image that are greater than 200.

[0051] In step S403 of some embodiments, the ratio of the number of pixel points in the first pixel group to the pixel points of the cell blood washing image is calculated to obtain the ratio of the pixel points in the first pixel group to the pixel points of the cell blood washing image, and a first ratio is obtained. Then, the ratio of the pixel points in the second pixel group to the pixel points of the cell blood washing image is obtained to obtain a second ratio.

[0052] In step S404 of some embodiments, the pixel variance value of each gray value is calculated according to the second pixel group corresponding to the first pixel group, the first ratio, and the second ratio. Among them, the aggregated calculation to obtain the pixel contrast difference is shown in formula (1): (1), Among them, is the pixel variance value, is the first ratio, is the second ratio, is the average gray value of the first pixel group, is the average gray value of the second pixel group.

[0053] In step S405 of some embodiments, the maximum pixel variance value is obtained, and the corresponding gray value threshold is obtained according to the maximum variance value to obtain the segmentation threshold, thereby completing the adaptive segmentation of the image without setting the segmentation threshold by oneself. In this embodiment, the optimal segmentation threshold is adaptively calculated from the pixel points of the image, improving the accuracy of image segmentation.

[0054] In step S406 of some embodiments, the cell blood washing image is binarized by the segmentation threshold. The pixel points in the cell blood washing image with pixel values greater than or equal to the segmentation threshold are set to 0, and the pixel points in the cell blood washing image with pixel values less than the segmentation threshold are set to 1 to obtain the binary cell image.

[0055] Please refer to Figure 5 , in some embodiments, step S203 includes but is not limited to steps S501 to S502: Step S501, image scaling is performed on the binary cell image according to a preset image scaling ratio to obtain a selected scaled image; Step S502, image cropping is performed on the selected scaled image according to a preset cropping window to obtain a selected cell image.

[0056] Steps S501 to S502 shown in the embodiments of the present application, by performing image scaling on the binary cell image to scale the image to a preset image size, and performing image cropping on the scaled image to obtain the valid area of the scaled image, without performing full-scale recognition of the image, improving the speed of evaluating the quality of the cell section image.

[0057] In step S501 of some embodiments, the image resolution of the binary cell image is the same as that of the cell section image, and the cell section image is the image after the thyroid section is scanned by the scanner. The cells are small in the binary cell image. Image magnification is performed on the binary cell image according to the image scaling ratio, facilitating subsequent image content detection and improving the detection accuracy of the thyroid cell information detection model for the selected cell image.

[0058] Please refer to Figure 10 , in step S502 of some embodiments, the cropping window is a cropping size of 1024*1024. First, the image is cropped according to the cropping window to obtain the first selected cell image, and it is judged whether the first selected cell image is a valid area. When the first selected cell image is not a valid area, the cropping window is continuously slid to crop the next selected cell image until the cropped selected cell image is a valid area.

[0059] In step S103 of some embodiments, the selected cell images include thyroid cell images and impurity images. The thyroid cell detection model is used to judge the selected cell images. First, the thyroid cell detection model performs image recognition to determine the cell mass images and impurity images in the selected cell images, then calculates the confidence of each cell mass image to obtain cell mass information, and calculates the confidence of the impurity images to obtain impurity information.

[0060] In one embodiment, the thyroid cell detection model includes a convolutional neural network and a multi-class neural network. The convolutional neural network is used to perform image recognition on the selected cell images to identify the positions of cell masses and impurities in the selected cell images, and input the positions of cell masses and impurities into the multi-class neural network to obtain the confidence of each cell mass and the confidence of each impurity.

[0061] In one embodiment, the thyroid cell detection model is a support vector machine model. The selected cell images are input into the support vector machine model. The support vector machine model classifies the selected cell images based on the pixel values of the selected cell images to obtain an image classification result, and calculates the confidence of the image classification result to obtain the confidence of each cell mass and the confidence of each impurity.

[0062] Please refer to Figure 6 , in some embodiments, the cell mass information includes thyroid follicular epithelial cell mass information and impurity cell mass information; step S104 includes but is not limited to steps S601 to S604: Step S601, obtain the confidence of the thyroid follicular epithelial cell mass information to obtain thyroid follicular epithelial cell mass confidence information; Step S602, screen the thyroid follicular epithelial cell mass information according to a preset confidence threshold and the thyroid follicular epithelial cell mass confidence information to obtain selected cell mass information; Step S603, input the selected cell mass information, impurity cell mass information and image impurity information into a slice quality evaluation model to count the number of cell masses to obtain the number of thyroid cell masses; Step S604, determine the thyroid cell quality evaluation result of the cell slice image according to a preset thyroid cell mass determination rule and the number of thyroid cell masses.

[0063] In the steps S601 to S604 illustrated in the embodiments of the present application, by obtaining the confidence level of the thyroid follicular epithelial cell mass information, the confidence level information of the thyroid follicular epithelial cell mass is obtained, and the thyroid follicular epithelial cell mass information is screened according to a preset confidence level threshold and the confidence level information of the thyroid follicular epithelial cell mass to obtain the selected cell mass information. Then, the selected cell mass information, the impurity cell mass information, and the image impurity information are input into the slice quality evaluation model for cell mass quantity statistics to obtain the thyroid cell mass quantity. Thus, by inputting the thyroid follicular epithelial cell mass, impurity cell mass information, and image impurity information with high confidence levels into the slice quality evaluation model for cell mass quantity statistics, the accuracy of the statistical quantity of the thyroid follicular epithelial cell mass is improved. Further, the thyroid cell quality evaluation result of the thyroid cell mass quantity is accurately determined according to the thyroid cell mass determination rule.

[0064] In step S601 of some embodiments, the evaluation data of the thyroid follicular epithelial cell mass information includes the confidence level information of the thyroid follicular epithelial cell mass and the position information of the thyroid follicular epithelial cell mass. The confidence level information of the thyroid follicular epithelial cell mass represents the probability that the cell mass is a thyroid follicular epithelial cell mass. The greater the confidence level, the greater the probability that the cell mass is a thyroid follicular epithelial cell mass.

[0065] In step S602 of some embodiments, the confidence level threshold is the top 10 thyroid follicular epithelial cell masses with the largest confidence levels. The top 10 confidence level information of the thyroid follicular epithelial cell masses is obtained from the multiple confidence level information of the thyroid follicular epithelial cell masses, and the corresponding thyroid follicular epithelial cell mass information is obtained to obtain the selected cell mass information. When the number of the confidence level information of the thyroid follicular epithelial cell masses is less than 10, zero-padding processing is performed to make the number of the confidence level information of the thyroid follicular epithelial cell masses reach 10. By screening the thyroid follicular epithelial cell mass information through the confidence level threshold, the input of interfering information is avoided, thereby improving the accuracy of quality evaluation.

[0066] It should be noted that the present application does not limit the value of the confidence level threshold, and it can be set to other values in other embodiments.

[0067] In step S603 of some embodiments, the impurity cell cluster information includes eosinophil cluster information, lymphocyte cluster information, macrophage cluster information, and multinucleated giant cell cluster information. In the cell section image, the eosinophil cluster information and lymphocyte cluster information have a positive effect on the thyroid follicular epithelial cell cluster, while the skeletal muscle cell information, ciliated cell information, macrophage cluster information, and multinucleated giant cell cluster information can be used as screening information to screen the cell clusters in the image. The image impurity information includes colloid information and cell degeneration product information. By inputting the selected cell cluster information, impurity cell cluster information, and image impurity information into the section quality assessment model for cell cluster quantity statistics, the input information volume is increased, so that the section quality assessment model can accurately perform quantity statistics on the thyroid follicular epithelial cell cluster and obtain the quantity of thyroid cell clusters.

[0068] In one embodiment, the section quality assessment model is a decision tree model. Input the selected cell cluster information, impurity cell cluster information, and image impurity information into the decision tree model, and obtain the quantity of thyroid cell clusters.

[0069] In one embodiment, the section quality assessment model is a DNN model. Obtain the quantity of thyroid cell clusters by inputting the selected cell cluster information, impurity cell cluster information, and image impurity information into the DNN model.

[0070] In step S604 of some embodiments, the thyroid cell cluster determination rules include a quantity determination rule and a quality determination rule. The thyroid cell cluster determination rule is that the set cell cluster threshold is 6 clusters. When the quantity of thyroid cell clusters is greater than or equal to 6 clusters, determine that the cell section image is a qualified quality section. The quality determination rule is that when the quantity of thyroid cell clusters is less than 6 clusters, obtain the colloid information from the image impurity information. If the colloid information in the image is greater than the preset colloid information threshold, determine that the cell section image is a qualified quality section. When both the quantity determination rule and the quality determination rule do not conform, determine that the cell section image is a non-qualified quality section.

[0071] Please refer to Figure 7 , in some embodiments, after step S104, it may further include but is not limited to steps S701 to S702: After inputting the cell cluster information and the image impurity information into a preset section quality assessment model for quality assessment to obtain the thyroid cell quality assessment result of the cell section image, it further includes: Step S701, if the thyroid cell quality assessment result indicates that the quantity of thyroid cell clusters conforms to the thyroid cell cluster determination rule, output section qualified information; Step S702: If the thyroid cell mass assessment result indicates that the number of thyroid cell clusters does not conform to the thyroid cell cluster determination rule, perform image sliding cropping on the selected scaled image according to the preset sliding window step size and cropping window to obtain an updated image; input the updated image into the thyroid cell information detection model for image content detection to obtain updated image content information; wherein, the updated image content information includes cell cluster update information and impurity update information; input the cell cluster update information and impurity update information into the slice quality assessment model to perform quality assessment on the updated image to obtain an updated cell mass assessment result.

[0072] For steps S701 to S702 illustrated in the embodiments of the present application, when the thyroid cell mass assessment result indicates that the number of thyroid cell clusters conforms to the thyroid cell cluster determination rule, output slice qualified information; when the thyroid cell mass assessment result indicates that the number of thyroid cell clusters does not conform to the thyroid cell cluster determination rule, perform image sliding cropping on the selected scaled image according to the preset sliding window step size and cropping window to obtain an updated image; input the updated image into the thyroid cell information detection model for image content detection to obtain updated image content information; wherein, the updated image content information includes cell cluster update information and impurity update information; input the cell cluster update information and impurity update information into the slice quality assessment model to perform quality assessment on the updated image to obtain an updated cell mass assessment result. In this way, it is not necessary to perform full-scale recognition on the cell slice image, and rapid quality assessment of the cell slice image can be achieved.

[0073] In step S701 of some embodiments, when the thyroid cell mass assessment result indicates that the number of thyroid cell clusters conforms to the thyroid cell cluster determination rule, output slice qualified information.

[0074] In step S702 of some embodiments, when the thyroid cell mass assessment result indicates that the number of thyroid cell clusters does not conform to the thyroid cell cluster determination rule, perform image sliding cropping on the selected scaled image according to the preset sliding window step size and cropping window to obtain an updated image; input the updated image into the thyroid cell information detection model for image content detection to obtain updated image content information; wherein, the updated image content information includes cell cluster update information and impurity update information; input the cell cluster update information and impurity update information into the slice quality assessment model to perform quality assessment on the updated image to obtain an updated cell mass assessment result.

[0075] In one embodiment, when the thyroid cell mass assessment result indicates that the number of thyroid cell clusters does not conform to the Susonghu thyroid cell cluster determination rule, the selected scaled image is subjected to image sliding cropping according to the sliding window step size and the cropping window to obtain an updated image, and the updated image is input into the thyroid cell information detection model for image content detection to obtain the confidence information of thyroid follicular epithelial cell clusters in the updated image. Then, the updated image and the corresponding confidence information of thyroid follicular epithelial cell clusters are added to a preset updated image set. Next, screening is performed according to the confidence information of each thyroid follicular epithelial cell cluster in the updated image set to obtain the top 5 updated images corresponding to the confidence information in the updated image set, thereby obtaining the selected updated images, and the quality of the selected updated images is evaluated. It should be noted that in this embodiment, the application does not specifically limit the acquisition of the top 5 updated images corresponding to the confidence information. The specific number of images acquired can be set to other values in other scenarios.

[0076] Please refer to Figure 11 , in the embodiment of the present application, after acquiring the cell section image, blood clearance processing is first performed, and the cell section image is cleared of blood from the aspects of color and morphology respectively to obtain a cell blood-washed image. Then, the cell blood-washed image is subjected to adaptive binarization processing to obtain the optimal segmentation threshold based on the pixel points of the image, and the cell blood-washed image is binarized according to the segmentation threshold to obtain a binarized cell image. Then, the binarized image is cropped to obtain a selected cell image, and the selected cell image is input into the thyroid cell detection model for image content detection to obtain the cell cluster information and impurity information in the image. Next, the cell cluster information and impurity information are input into the slice quality assessment model for quality assessment to obtain a quality assessment result, which provides a large amount of image information for the slice quality assessment model and improves the accuracy of thyroid cell cluster recognition. When the quality assessment result indicates that the number of thyroid cell clusters in the selected cell image is greater than or equal to the cell cluster threshold, slice qualified information is output. When the quality assessment result indicates that the number of thyroid cell clusters in the selected cell image is less than the cell cluster threshold, it is determined whether it is the last image block in the binarized image. When it is not the last image block, the next selected cell image is obtained according to the sliding window step size and the cropping window, thereby avoiding full-scale recognition of the image and accelerating the speed of image quality assessment. When it is the last image block, the entire quality assessment process ends.

[0077] Please refer to Figure 8 , the embodiment of the present application further provides a thyroid slice image quality assessment device, which can implement the above-mentioned thyroid slice image quality assessment method. The device includes: An acquisition module 801, configured to acquire a cell section image; wherein, the cell section image is obtained by scanning a thyroid cell section by a scanner; An extraction module 802, configured to extract cell images from the cell section images to obtain selected cell images; A detection module 803, configured to input the selected cell images into a preset thyroid cell information detection model for image content detection to obtain cell image content information; wherein, the cell image content information includes: cell cluster information and image impurity information; An evaluation module 804, configured to input the cell cluster information and the image impurity information into a preset section quality evaluation model for quality evaluation to obtain a thyroid cell quality evaluation result of the cell section images.

[0078] The specific implementation manner of the thyroid section image quality evaluation device is basically the same as that of the specific embodiments of the above-mentioned thyroid section image quality evaluation method, and will not be elaborated herein.

[0079] An embodiment of the present application further provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above-mentioned thyroid section image quality evaluation method is implemented. The electronic device can be any intelligent terminal including a tablet computer, a vehicle-mounted computer, etc.

[0080] Please refer to Figure 9 , Figure 9 , which schematically shows the hardware structure of an electronic device in another embodiment. The electronic device includes: A processor 901, which can be implemented in a general-purpose CPU (Central Processing Unit, central processor), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, etc., and is configured to execute relevant programs to implement the technical solutions provided by the embodiments of the present application; A memory 902, which can be implemented in the form of a read-only memory (Read Only Memory, ROM), a static storage device, a dynamic storage device, or a random access memory (Random Access Memory, RAM), etc. The memory 902 can store an operating system and other application programs. When implementing the technical solutions provided by the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 902, and the processor 901 is called to execute the thyroid section image quality evaluation method of the embodiments of the present application; An input / output interface 903, configured to implement information input and output; A communication interface 904, configured to implement communication interaction between this device and other devices, and can implement communication through a wired manner (such as USB, network cable, etc.) or through a wireless manner (such as a mobile network, WIFI, Bluetooth, etc.); A bus 905 transmits information between various components of the device, such as a processor 901, a memory 902, an input / output interface 903, and a communication interface 904; Among them, the processor 901, the memory 902, the input / output interface 903, and the communication interface 904 are communicatively connected to each other inside the device through the bus 905.

[0081] An embodiment of the present application also provides a computer-readable storage medium storing a computer program, which when executed by a processor implements the above-mentioned thyroid slice image quality assessment method.

[0082] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory may include high-speed random access memory and may also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory may optionally include a memory remotely located relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above networks include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0083] The thyroid slice image quality assessment method, thyroid slice image quality assessment device, device, and storage medium provided by the embodiments of the present application extract cell images from cell slice images obtained by scanning thyroid cell slices by a scanner, extract image content related to thyroid cells from the cell slice images to obtain selected cell images, so as to exclude interference information in the images, avoid interference of the interference information in the images on subsequent image processing, and input the selected cell images into a thyroid cell information detection model for image content detection to obtain cell cluster information and image impurity information, and then input the cell cluster information and the image impurity information into a slice quality assessment model for quality assessment, the thyroid cell quality assessment result, input the image impurity information and the cell cluster information into the slice quality assessment model at the same time, and combine the cell impurity information and the cell cluster information to perform quality assessment on the thyroid cell slice, which increases the amount of information input to the slice quality model, thereby improving the quality assessment accuracy of the slice quality assessment model for cell slice images, that is, the thyroid slice image quality assessment method provided by the present disclosure can improve the quality assessment accuracy of scanned slice images and obtain high-quality thyroid cell slice images.

[0084] The embodiments described in the embodiments of the present application are to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art will know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.

[0085] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than those shown, or combine certain steps, or different steps.

[0086] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0087] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and their appropriate combinations.

[0088] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0089] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or similar expressions refer to any combination of these items, including any combination of single items (ones) or plural items (ones). For example, at least one (one) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0090] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the above-mentioned division of units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be in electrical, mechanical or other forms.

[0091] The units described above as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they can be located in one place, or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0092] In addition, in each embodiment of this application, each functional unit can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0093] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of various embodiments of this application. The aforementioned storage medium includes: various media that can store programs such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0094] The preferred embodiments of the embodiments of this application have been described above with reference to the accompanying drawings, and thus do not limit the scope of the rights of the embodiments of this application. Any modifications, equivalent replacements, and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of this application shall be within the scope of the rights of the embodiments of this application.

Claims

1. A method for evaluating the quality of a thyroid slice image, characterized in that: The method comprises: Acquire a cell slice image; wherein the cell slice image is obtained by scanning a thyroid cell slice with a scanner; Performing cell image extraction on the cell slice image to obtain a selected cell image; Inputting the selected cell image into a preset thyroid cell information detection model to perform image content detection to obtain cell image content information; wherein the cell image content information includes: cell cluster information and image impurity information; The cell cluster information and the image impurity information are input into a preset slice quality assessment model to perform quality assessment to obtain a thyroid cell quality assessment result of the cell slice image.

2. The method according to claim 1, characterized in that The cell cluster information includes thyroid follicular epithelial cell cluster information and impurity cell cluster information; the cell cluster information and the image impurity information are input into a preset slice quality assessment model for quality assessment to obtain a thyroid cell quality assessment result of the cell slice image, including: Acquiring the confidence of the thyroid follicular epithelial cell cluster information to obtain the thyroid follicular epithelial cell cluster confidence information; Screening the thyroid follicular epithelial cell cluster information according to a preset confidence threshold and the thyroid follicular epithelial cell cluster confidence information to obtain selected cell cluster information; Inputting the selected cell cluster information, the impurity cell cluster information and the image impurity information into the slice quality assessment model to count the number of cell clusters to obtain the number of thyroid cell clusters; The thyroid cell quality assessment result of the cell slice image is determined according to a preset thyroid cell cluster determination rule and the number of the thyroid cell clusters.

3. The method according to claim 1, characterized in that The step of extracting a cell image from the cell slice image to obtain a selected cell image comprises: Cleaning the blood area in the cell slice image using a preset image blood area cleaning model to obtain a cell blood cleaned image; Binarizing and segmenting the cell blood cleansing image to obtain a binary cell image; The binary cell image is cropped to obtain a selected cell image.

4. The method according to claim 3, characterized in that The step of performing binary segmentation on the cell blood cleansing image to obtain a binary cell image comprises: Obtaining the number of pixels of the cell blood cleaned image, and obtaining the pixel gray value of each pixel in the cell blood cleaned image; Classifying the plurality of pixel grayscale values ​​according to a plurality of preset grayscale value thresholds to obtain pixel classification group pairs corresponding to each grayscale value threshold; wherein the pixel classification group pairs include a first pixel group and a second pixel group; Obtaining a ratio of each of the first pixel groups to the number of pixel points to obtain a first ratio, and obtaining a ratio of each of the second pixel groups to the number of pixel points to obtain a second ratio; Performing aggregation calculation on the second pixel group, the first ratio, and the second ratio corresponding to each of the first pixel groups to obtain a pixel variance value of each gray value threshold; Obtaining the maximum pixel variance value to obtain a segmentation variance value, and obtaining the corresponding gray value threshold according to the segmentation variance value to obtain a segmentation threshold; The cell blood cleansing image is subjected to image binarization processing according to the segmentation threshold value to obtain the binarized cell image.

5. The method according to claim 3, characterized in that: The image blood region cleaning model includes a color cleaning sub-model and a morphology cleaning sub-model; the blood region in the cell slice image is cleaned by the preset image blood region cleaning model to obtain a cell blood cleaned image, including: Performing color screening on the cell slice image by using the color cleaning sub-model to obtain a selected color image; wherein the selected color image represents an image without blood color; The selected color image is subjected to morphological screening by the morphological cleaning sub-model to obtain the cell blood cleaned image; wherein the cell blood cleaned image represents an image without blood morphology.

6. The method according to claim 3, characterized in that The step of performing image cropping on the binary cell image to obtain a selected cell image comprises: Scaling the binary cell image according to a preset image scaling ratio to obtain a selected scaled image; The selected zoomed image is cropped according to a preset cropping window to obtain the selected cell image.

7. The method according to any one of claims 1 to 6, characterized in that: After the cell cluster information and the image impurity information are input into a preset slice quality assessment model for quality assessment to obtain a thyroid cell quality assessment result of the cell slice image, the method further includes: If the thyroid cell quality assessment result indicates that the number of thyroid cell clusters meets the thyroid cell cluster determination rule, outputting slice qualification information; If the thyroid cell quality assessment result is characterized by the number of thyroid cell clusters not meeting the thyroid cell cluster determination rule, the selected zoomed image is subjected to image sliding cropping according to a preset sliding window step size and the cropping window to obtain an updated image; the updated image is input into the thyroid cell information detection model for image content detection to obtain updated image content information; wherein the updated image content information includes cell cluster update information and impurity update information; the cell cluster update information and the impurity update information are input into the slice quality assessment model to perform quality assessment on the updated image to obtain an updated cell quality assessment result.

8. A thyroid slice image quality assessment device, characterized in that: The device comprises: An acquisition module, used for acquiring a cell slice image; wherein the cell slice image is obtained by scanning a thyroid cell slice with a scanner; An extraction module, used for performing cell image extraction on the cell slice image to obtain a selected cell image; A detection module, used for inputting the selected cell image into a preset thyroid cell information detection model to perform image content detection to obtain cell image content information; wherein the cell image content information includes: cell cluster information and image impurity information; The evaluation module is used to input the cell cluster information and the image impurity information into a preset slice quality evaluation model to perform quality evaluation and obtain a thyroid cell quality evaluation result of the cell slice image.

9. An electronic device, characterized in that: The electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the thyroid slice image quality assessment method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for evaluating the quality of thyroid slice images according to any one of claims 1 to 7 is implemented.

Images