Medical image quality control method and computer readable storage medium

Abstract

The present application relates to a quality control method for medical images and a computer-readable storage medium. The method includes: obtaining an X-ray knee joint image, the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image; inputting the X-ray knee joint image into an image segmentation model to obtain an image segmentation result; the image segmentation result includes an anteroposterior image segmentation result and / or a lateral image segmentation result, and the image segmentation result includes at least one of a femur mask, a fibula mask, a tibia mask, a patella mask, and a knee joint gap mask; based on the image segmentation result, calculating the quantitative result of the X-ray knee joint image under a first quality index; the first quality index is used to characterize whether the shooting position of each tissue in the X-ray knee joint image is standardized. The method greatly improves the efficiency of the quality control process of medical images; at the same time, it greatly improves the accuracy of the image segmentation results obtained, thereby improving the accuracy of the quality quantification results of the X-ray knee joint image.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a quality control method for medical images and a computer-readable storage medium. Background Art

[0002] With the continuous development of the medical industry, the continuous improvement of the social medical system and the increasing living standards of the people, the number of daily examinations in hospitals has gradually increased. Due to the relatively low price and relatively quick examination steps of digital radiography (DR) equipment, the number of X-ray examinations is also in a rapid growth stage. In knee joint X-ray examinations, due to the limited number of shooting technicians and the uneven operation level, the poor quality of knee joint images frequently occurs, which affects the doctor's later diagnosis; and currently, difficult cases shot by lower-level hospitals will be uniformly uploaded to higher-level hospitals. If the image quality of the lower-level hospital is poor and affects the diagnosis, the patient will need to take a second shot, resulting in a waste of medical resources.

[0003] In traditional technology, manual methods are used to perform random inspections on knee joint images. Generally, relevant elements are checked through visual inspection or simple measurement, such as: whether the basic information of the image is complete, whether the left and right markings are correct, whether there are foreign objects and artifacts in the image, whether the exposure conditions are appropriate, whether the diagnostic area is clearly visible, etc., and then the inspection results are sent to the corresponding technicians for improvement.

[0004] However, due to the increased workload of filming, it is difficult for technicians to have time to conduct detailed analysis of the filmed films and control their filming quality. Therefore, the medical image quality control results in traditional technologies are less efficient. Summary of the invention

[0005] Based on this, it is necessary to provide a medical image quality control method and a computer-readable storage medium to address the problem of low efficiency of medical image quality control results in traditional technologies.

[0006] A quality control method for medical images, the method comprising:

[0007] Acquire an X-ray knee joint image, wherein the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image;

[0008] Inputting the X-ray knee joint image into the image segmentation model to obtain an image segmentation result; the image segmentation result includes an anteroposterior image segmentation result and / or a lateral image segmentation result, and the image segmentation result includes at least one of a femur mask, a fibula mask, a tibia mask, a patella mask, and a knee joint space mask;

[0009] Based on the image segmentation result, a quantitative result of the X-ray knee joint image under a first quality index is calculated; the first quality index is used to characterize whether the shooting position of each tissue in the X-ray knee joint image is standardized.

[0010] In one embodiment, the first quality indicator includes a first sub-indicator and / or a second sub-indicator; based on the image segmentation result, calculating the quantitative result of the X-ray knee joint image under the first quality indicator includes:

[0011] Based on the segmentation result of the anteroposterior image, a quantitative result of the anteroposterior knee joint image under the first sub-index is calculated; the first sub-index includes at least one of a joint long axis position index, a femoral intercondylar fossa position index, a fibula and tibia overlap index, a patella position index, and a knee joint space size index;

[0012] and / or,

[0013] Based on the lateral image segmentation result, the quantitative result of the lateral knee joint image under the second sub-indicator is calculated; the second sub-indicator includes at least one of the tibia and fibula long axis position index, the patellar lower edge position index, the femur and patella overlap index, the femur and fibula overlap index and the femur and fibula angle index.

[0014] In one embodiment, based on the segmentation result of the anteroposterior image, calculating the quantitative result of the anteroposterior knee joint image under the first sub-index includes at least one of the following steps:

[0015] Based on the long axis direction of the femoral mask in the segmentation result of the anteroposterior image and the main axis direction of the anteroposterior knee joint image, the quantitative result under the long axis position index of the joint is determined;

[0016] Based on the center point position of the lowest edge of the femoral mask in the segmentation result of the anteroposterior image, the quantitative result under the femoral intercondylar fossa position index is determined;

[0017] Based on the overlapping area of ​​the fibula mask and the tibia mask in the segmentation result of the orthotopic image, the quantitative result under the overlap index of the fibula and the tibia is determined;

[0018] Based on the left and right boundaries of the patella mask and the left and right boundaries of the femur mask in the segmentation result of the frontal image, the quantitative result under the patella position index is determined;

[0019] Based on the left and right gaps of the knee joint gap mask in the segmentation results of the anteroposterior image, the quantitative results under the knee joint gap size index are determined.

[0020] In one embodiment, based on the lateral image segmentation result, calculating the quantitative result of the lateral knee joint image under the second sub-index includes at least one of the following steps:

[0021] Based on the long axis directions of the tibia mask and the fibula mask in the lateral image segmentation result and the main axis direction of the lateral knee joint image, the quantitative results under the long axis position index of the tibia and fibula are determined;

[0022] Based on the center point position of the lowest edge of the patellar mask in the lateral image segmentation result, the quantitative result under the patellar lower edge position index is determined;

[0023] Based on the overlapping area of ​​the femur mask and the patella mask in the lateral image segmentation result, the quantitative result under the femur and patella overlap index is determined;

[0024] Based on the overlapping area of ​​the femur mask and the fibula mask in the lateral image segmentation result, the quantitative result under the femur and fibula overlap index is determined;

[0025] Based on the long axis direction of the femur mask and the long axis direction of the fibula mask in the lateral image segmentation result, the quantitative result under the femur-fibula angle index is determined.

[0026] In one embodiment, the method further comprises:

[0027] Inputting the X-ray knee joint image into the image classification model to obtain an image classification result; the image classification result includes at least one of a grayscale density classification result, an artifact classification result, and a clarity classification result;

[0028] Based on the image classification result, a quantitative result of the X-ray knee joint image under a second quality index is calculated; the second quality index is used to characterize whether the X-ray knee joint image is clear.

[0029] In one embodiment, the second quality index includes at least one of an image density index, an artifact index, and an image clarity index; and based on the image classification result, calculating the quantitative result of the X-ray knee joint image under the second quality index includes at least one of the following steps:

[0030] Determine a quantization result under the image density index based on the grayscale density classification result and a preset first threshold value;

[0031] Determine a quantification result under the artifact index based on the artifact classification result and a preset second threshold;

[0032] Based on the clarity classification result and a preset third threshold, a quantization result under the image clarity index is determined.

[0033] In one embodiment, the method further comprises:

[0034] Inputting the X-ray knee joint image into the image detection model to obtain an image detection result; the image detection result includes a foreign body detection result and / or an image marker detection result;

[0035] Based on the image detection result, the quantitative result of the X-ray knee joint image under the third quality index is calculated; the third quality index is used to characterize the type of non-tissue part in the X-ray knee joint image.

[0036] In one embodiment, the third quality indicator includes a foreign body indicator and / or an image marking indicator; based on the image detection result, the quantitative result of the X-ray knee joint image under the third quality indicator is calculated, including:

[0037] Based on the foreign matter detection result and a preset fourth threshold, determining a quantitative result under the foreign matter index;

[0038] and / or,

[0039] Based on the image marker detection results and the image labels of the X-ray knee joint images, the quantitative results under the image marker indicators are determined.

[0040] In one embodiment, the method further comprises:

[0041] The quantification results are sent to the client of the corresponding user in real time through the network transmission protocol, so that the user can choose whether to re-collect the X-ray knee joint image according to the quantification results;

[0042] and / or,

[0043] Store X-ray knee joint images and quantitative results in a database;

[0044] Taking a preset time interval as a period, statistics are performed on the X-ray knee joint images and quantification results in the database to obtain statistical results; the statistical results represent the quality changes of the X-ray knee joint images collected by the user in the current period;

[0045] The statistical results are displayed to the user's client via a web page.

[0046] A quality control device for medical images, comprising:

[0047] An acquisition module, used for acquiring an X-ray knee joint image, wherein the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image;

[0048] An image segmentation module, used for inputting an X-ray knee joint image into an image segmentation model to obtain an image segmentation result; the image segmentation result includes an anteroposterior image segmentation result and / or a lateral image segmentation result, and the image segmentation result includes at least one of a femur mask, a fibula mask, a tibia mask, a patella mask, and a knee joint space mask;

[0049] The quantification module is used to calculate the quantification result of the X-ray knee joint image under the first quality index based on the image segmentation result; the first quality index is used to characterize whether the shooting position of each tissue in the X-ray knee joint image is standardized.

[0050] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0051] Acquire an X-ray knee joint image, wherein the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image;

[0052] Inputting the X-ray knee joint image into the image segmentation model to obtain an image segmentation result; the image segmentation result includes an anteroposterior image segmentation result and / or a lateral image segmentation result, and the image segmentation result includes at least one of a femur mask, a fibula mask, a tibia mask, a patella mask, and a knee joint space mask;

[0053] Based on the image segmentation result, a quantitative result of the X-ray knee joint image under a first quality index is calculated; the first quality index is used to characterize whether the shooting position of each tissue in the X-ray knee joint image is standardized.

[0054] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0055] Acquire an X-ray knee joint image, wherein the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image;

[0056] Inputting the X-ray knee joint image into the image segmentation model to obtain an image segmentation result; the image segmentation result includes an anteroposterior image segmentation result and / or a lateral image segmentation result, and the image segmentation result includes at least one of a femur mask, a fibula mask, a tibia mask, a patella mask, and a knee joint space mask;

[0057] Based on the image segmentation result, a quantitative result of the X-ray knee joint image under a first quality index is calculated; the first quality index is used to characterize whether the shooting position of each tissue in the X-ray knee joint image is standardized.

[0058] The quality control method, device, computer equipment and storage medium of the above-mentioned medical image can obtain an X-ray knee joint image, which includes an anteroposterior knee joint image and / or a lateral knee joint image; the X-ray knee joint image is input into an image segmentation model to obtain an image segmentation result; the image segmentation result includes an anteroposterior image segmentation result and / or a lateral image segmentation result, and the image segmentation result includes at least one of a femur mask, a fibula mask, a tibia mask, a patella mask and a knee joint gap mask; based on the image segmentation result, the quantification result of the X-ray knee joint image under the first quality index is calculated; the first quality index is used to characterize whether the shooting position of each tissue in the X-ray knee joint image is standardized. In this method, the X-ray knee joint image is automatically obtained by a computer device, and the X-ray knee joint image is segmented and quantified to obtain the corresponding quality quantification result. The entire data processing process does not require human participation, which greatly improves the efficiency of the quality control process of the medical image; at the same time, the image segmentation is performed by the computer device through the image segmentation model, which also greatly improves the accuracy of the obtained image segmentation result, thereby improving the accuracy of the quality quantification result of the X-ray knee joint image. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is an internal structure diagram of a computer device in one embodiment;

[0060] Figure 2 is a schematic diagram of a process of a quality control method for medical images in one embodiment;

[0061] Figure 2a A schematic diagram of a segmentation result of a frontal image and a segmentation result of a lateral image in one embodiment;

[0062] Figure 3 is a schematic flow chart of a quality control method for medical images in another embodiment;

[0063] Figure 4 is a schematic flow chart of a quality control method for medical images in yet another embodiment;

[0064] Figure 5 is a schematic flow chart of a quality control method for medical images in yet another embodiment;

[0065] Figure 6 is a schematic flow chart of a quality control method for medical images in yet another embodiment;

[0066] Figure 7 FIG. 4 is a structural block diagram of a quality control device for medical images in one embodiment. DETAILED DESCRIPTION

[0067] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with 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.

[0068] The quality control method of medical images provided in the embodiments of the present application can be applied to Figure 1 The computer device shown. The computer device includes a processor and a memory connected by a system bus, and a computer program is stored in the memory. When the processor executes the computer program, the steps of the following method embodiment can be executed. Optionally, the computer device may also include a communication interface, a display screen and an input device. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory, and the non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal through a network connection. Optionally, the computer device can be a personal computer (personal computer, referred to as PC), a personal digital assistant, or other terminal devices, such as a tablet computer (portable android device, referred to as PAD), a mobile phone, etc., or a cloud or remote server. The embodiment of the present application does not limit the specific form of the computer device.

[0069] In one embodiment, Figure 2 As shown, a method for quality control of medical images is provided. This embodiment relates to a specific process of segmenting an X-ray knee joint image to determine its quantitative result under a first quality index. Figure 1 The computer device in the example is used to illustrate, including the following steps:

[0070] S101, acquiring an X-ray knee joint image, wherein the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image.

[0071] Specifically, the X-ray knee joint image can be obtained from a Picture Archiving and Communication Systems (PACS) by a computer device, and the X-ray knee joint image can include an anteroposterior knee joint image and / or a lateral knee joint image. Optionally, after obtaining the X-ray knee joint image, the computer device can also perform preprocessing operations on it, including but not limited to: processing the X-ray knee joint image into a size acceptable to the following image segmentation model, normalizing the image through a predetermined window width and window position within the image, and standardizing the image according to a set mean value and variance, etc.

[0072] S102, input the X-ray knee joint image into the image segmentation model to obtain an image segmentation result; the image segmentation result includes a frontal image segmentation result and / or a lateral image segmentation result, and the image segmentation result includes at least one of a femur mask, a fibula mask, a tibia mask, a patella mask and a knee joint space mask.

[0073] Specifically, the computer device can input the above-mentioned preprocessed X-ray knee joint image into the image segmentation model, and use the image segmentation model to segment the femur, fibula, tibia, patella and the area where the knee joint gap is located in the X-ray knee joint image to obtain an image segmentation result. Among them, because the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image, the corresponding image segmentation result also includes the corresponding anteroposterior image segmentation result and / or lateral image segmentation result. Through the segmentation operation of the image segmentation model, at least one of the femur mask, fibula mask, tibia mask, patella mask and knee joint gap mask can be obtained, and the mask refers to a binary image that separates the target area from the background area. For schematic diagrams of the anteroposterior image segmentation results and the lateral image segmentation results, please refer to Figure 2a shown.

[0074] Optionally, the above-mentioned image segmentation model can be a neural network model, such as a convolutional neural network, a recurrent neural network, etc.; optionally, the image segmentation model includes but is not limited to network models such as Unet, Linknet, Deeplab, etc. Among them, the training method of the image segmentation model may include: obtaining a large number of sample images that meet the quality assessment specifications and do not meet the quality specifications, and having experienced doctors segment and annotate each sample image, that is, annotate the areas where the femur, fibula, tibia, patella and knee joint gap are located as the segmentation gold standard; then the computer device inputs the sample image into the initial image segmentation model, obtains the prediction mask of each area, and calculates the loss loss between the prediction mask and the corresponding segmentation gold standard, and updates the network parameters (weights) of the initial image segmentation model by back-propagating the loss; after continuous iteration, a training converged image segmentation model can be obtained. Optionally, in order to improve the robustness of the image segmentation model, the computer device can also perform data amplification on the sample image, and the data amplification method includes but is not limited to image rotation, flipping, contrast adjustment and brightness adjustment.

[0075] S103, based on the image segmentation result, calculating the quantitative result of the X-ray knee joint image under the first quality index; the first quality index is used to indicate whether the shooting position of each tissue in the X-ray knee joint image is standard.

[0076] Specifically, based on the above image segmentation results, the computer device can compare the relative positions between the masks to calculate the quantitative results of the X-ray knee joint image under the first quality index, and the first quality index is used to characterize whether the shooting positions of each tissue in the X-ray knee joint image are standardized. Exemplarily, for the anteroposterior knee joint image, the relative position of the femur mask in the entire X-ray knee joint image can be used to determine whether the shooting position of the knee joint is centered; for the lateral knee joint image, the relative position of the tibia mask in this X-ray knee joint image can be used to determine whether the shooting position of the knee joint is centered, and so on; if it is centered, the quantitative result is set to 1, and if it is not centered, it is set to 0.

[0077] Optionally, after acquiring the above-mentioned sample image, in addition to the segmentation gold standard annotation, the computer device can also perform quantitative numerical annotation on it under the first quality index, wherein the quantitative numerical annotation of the anteroposterior knee joint image includes the angle between the long axis of the joint and the long axis of the image, the judgment classification result that the femoral intercondylar fossa is located at the center of the image, the mask overlap ratio of the fibula and the tibia, the judgment classification result that the patella is located between the inner and outer malleolus, and the judgment classification result of the equidistant knee joint space; the quantitative numerical annotation of the lateral knee joint image includes the angle between the long axis of the tibia and the fibula and the long axis of the image, the judgment classification result that the lower edge of the patella is located at the center of the image, the judgment classification result of whether the femur and the patella overlap, the mask overlap ratio of the femur and the fibula, and the angle between the femur and the fibula. Exemplarily, the quantitative numerical annotation can be set to: the angle range is 0-180°, the mask overlap ratio range is 0-1, where 0 is no overlap, 1 is full overlap, and the judgment classification results are 0 and 1, where 0 means no and 1 means yes.

[0078] Then, a multi-task processing model is trained based on the sample image, the segmentation gold standard, and the quantitative numerical annotations: loss1 is calculated based on the mask output by the model and the segmentation gold standard, loss2 is calculated based on the quantitative data output by the model and the quantitative numerical annotations, and the multi-task processing model is updated after weighted summation of loss1 and loss2. Optionally, the loss function used in this embodiment includes but is not limited to the Dice function, the L2 function, the BCE function, etc. Based on the multi-task processing model with converged training, the computer device can directly input the X-ray knee joint image into the model. After obtaining the features of the image, the result can be divided into multiple branches, and the image segmentation result and the quantitative data can be calculated at the same time, that is, the image segmentation result of the X-ray knee joint image and its quantitative result under the first quality indicator are obtained.

[0079] The quality control method of medical images provided in this embodiment is that the computer device inputs the acquired X-ray knee joint image into the image segmentation model to obtain the image segmentation result, and the image segmentation result includes at least one of the femur mask, the fibula mask, the tibia mask, the patella mask and the knee joint gap mask; based on the image segmentation result, the quantification result of the X-ray knee joint image under the first quality index is calculated, and the first quality index is used to characterize whether the shooting position of each tissue in the X-ray knee joint image is standardized. In this method, the X-ray knee joint image is automatically acquired by the computer device, and the segmentation and quantification are performed on it to obtain the corresponding quality quantification result. The entire data processing process does not require human participation, which greatly improves the efficiency of the quality control process of medical images; at the same time, the image segmentation performed by the computer device through the image segmentation model also greatly improves the accuracy of the obtained image segmentation result, thereby improving the accuracy of the quality quantification result of the X-ray knee joint image.

[0080] In one embodiment, since the image segmentation result includes the anteroposterior image segmentation result and / or the lateral image segmentation result, the quantization angles of the anteroposterior knee joint image and the lateral knee joint image are different when performing quality quantization, the first quality index includes the first index and / or the second index. Optionally, Figure 3 As shown, the above S103 may include S201 and / or S202:

[0081] S201, based on the segmentation result of the AP image, calculate the quantitative result of the AP knee joint image under the first sub-indicator; the first sub-indicator includes at least one of the joint long axis position index, the femoral intercondylar fossa position index, the fibula and tibia overlap index, the patella position index and the knee joint space size index.

[0082] Specifically, based on the segmentation results of the orthogonal image (such as Figure 2a ), the computer device can calculate the quantitative result of the anteroposterior knee joint image under the first sub-index. Optionally, based on the femoral mask in the segmentation result of the anteroposterior image, its long axis direction can be obtained, and the angle value can be obtained by comparing with the main axis direction of the anteroposterior knee joint image. If the angle value is greater than a set threshold (such as 5°), it is considered that the joint long axis of the anteroposterior knee joint image does not coincide with the image long axis, otherwise they coincide, that is, the quantitative result under the joint long axis position index is determined. Based on the femoral mask in the segmentation result of the anteroposterior image, the distance between the center point position of the lowermost edge of the mask and each boundary of the image can be calculated to obtain the distance ratio. If the left-right ratio or the upper-lower ratio difference is less than a set threshold (such as 10mm), it is considered that the femoral intercondylar fossa of the anteroposterior knee joint image is located at the center of the image, otherwise it is not located at the center, that is, the quantitative result under the femoral intercondylar fossa position index is determined. Based on the fibula mask and tibia mask in the segmentation result of the orthotopic image, the overlapping area between the two can be calculated, and the overlapping area is divided by the sum of the areas of the fibula mask and the tibia mask to obtain the area ratio. If the area ratio is less than the set threshold (such as 5mm 2), then the overlapping part of the fibula and tibia meets the specification, otherwise it does not meet the specification, that is, the quantitative result under the fibula and tibia overlap index is determined. Based on the distance between the left and right boundaries of the patella mask and the left and right boundaries of the femur mask in the orthotopic image segmentation result, if the difference in distance is less than the set threshold (such as 5mm), it is considered that the patella is located between the inner and outer malleolus, otherwise it is not, that is, the quantitative result under the patella position index is determined. Based on the height mean of the left and right gaps of the knee joint gap mask in the orthotopic image segmentation result, if the height mean is less than the set threshold (such as 3mm), it is considered that the inner and outer gaps of the knee joint are equidistant, otherwise it does not meet the specification, that is, the quantitative result under the knee joint size index is determined. It should be noted that the quantitative results under the joint long axis position index, the femoral intercondylar fossa position index, the fibula and tibia overlap index, the patella position index and the knee joint gap size index in the above-mentioned first sub-indicator can be calculated according to at least one of the actual needs.

[0083] S202, based on the lateral image segmentation result, calculate the quantitative result of the lateral knee joint image under the second sub-indicator; the second sub-indicator includes at least one of the tibia and fibula long axis position index, the patellar lower edge position index, the femur and patella overlap index, the femur and fibula overlap index and the femur and fibula angle index.

[0084] Specifically, based on the lateral image segmentation results (such as Figure 2a ), the computer device can calculate the quantitative result of the lateral knee joint image under the second sub-index. Optionally, the computer device can compare the long axis direction of the area after the tibia mask and the fibula mask are merged in the lateral image segmentation result with the main axis direction of the lateral knee joint image to obtain an angle value. If the angle value is greater than a set threshold (such as 5°), it is considered that the long axis of the tibia and fibula of the image does not coincide with the long axis of the image, otherwise they coincide, that is, the quantitative result under the tibia and fibula long axis position indicator is determined. Based on the distance between the center point of the lower edge of the patella mask in the lateral image segmentation result and each boundary of the image, a distance ratio is obtained. If the left-right ratio or the upper-lower ratio difference is less than a predetermined threshold (such as 10 mm), it is considered that the lower edge of the patella of the lateral knee joint image is located in the center of the image, otherwise it is not located in the center of the image, that is, the quantitative result under the lower edge position indicator of the patella is obtained. Based on the femur mask and patella mask in the lateral image segmentation result, the overlapping area between the two can be calculated, and the overlapping area is divided by the sum of the areas of the femur mask and the patella mask to obtain the area ratio. If the area ratio is less than the set threshold (such as 5mm 2), the overlapping part of the femur and patella is considered to meet the specifications, otherwise it does not meet the specifications, that is, the quantitative result under the femur and patella overlap index is determined. Based on the femur mask and the fibula mask in the lateral image segmentation result, the overlapping area between the two can be calculated, and the overlapping area is divided by the sum of the areas of the femur mask and the fibula mask to obtain the area ratio. If the area ratio is less than the set threshold, the overlapping part of the femur and fibula is considered to meet the specifications, otherwise it does not meet the specifications, that is, the quantitative result under the femur and fibula overlap index is determined. Based on the long axis direction of the femur mask and the long axis direction of the fibula mask in the lateral image segmentation result, the angle value between the two is calculated. If the angle value is between the set range (such as 3°-5°), it is considered that the angle between the femur and fibula of the image meets the specifications, otherwise it does not meet the specifications, that is, the quantitative result under the tibia and fibula long axis position index is determined. It should be noted that, among the above-mentioned second sub-indicators, at least one of the quantitative results of the tibia and fibula long axis position index, the patellar lower edge position index, the femur and patella overlap index, the femur and fibula overlap index and the femur and fibula angle index can be calculated according to actual needs.

[0085] In the quality control method for medical images provided in this embodiment, the computer device can calculate the quantitative result of the anteroposterior knee joint image under the first sub-indicator based on the segmentation result of the anteroposterior image, and calculate the quantitative result of the lateral knee joint image under the second sub-indicator based on the segmentation result of the lateral image, wherein the first sub-indicator and the second sub-indicator are set according to the image characteristics of the anteroposterior knee joint image and the lateral knee joint image, and the obtained quantitative result also conforms to the quality characteristics of the actual X-ray knee joint image. Therefore, this embodiment can greatly improve the quality control efficiency of medical images under the premise of ensuring that the obtained quality quantification result meets the actual needs.

[0086] In one embodiment, the computer device can perform image classification on the X-ray knee joint image in addition to image segmentation. Figure 4 As shown, the above method also includes:

[0087] S301, inputting the X-ray knee joint image into an image classification model to obtain an image classification result; the image classification result includes at least one of a grayscale density classification result, an artifact classification result and a clarity classification result.

[0088] Specifically, the computer device can input the above X-ray knee joint image into the image classification model, and determine whether the grayscale density of the X-ray knee joint image is normal, whether there are artifacts, and whether the clarity is normal through the image classification model, so as to obtain at least one of the grayscale density classification result, the artifact classification result, and the clarity classification result. Optionally, the image classification result can be expressed in the form of scores (or probabilities) in each classification, such as the grayscale density classification result is 90 points, the artifact classification result is 80 points, the clarity classification result is 50 points, etc.

[0089] Optionally, the above-mentioned image classification model can be a neural network model, such as a convolutional neural network, a recurrent neural network, etc.; optionally, the image classification model includes but is not limited to Ensenet, Resnet, and Senet models. The training method of the image classification model may include: obtaining a large number of training images, and using the above-mentioned preprocessing method and data augmentation method to process the training images to obtain the final sample image, and annotating the classification results of whether the grayscale density of the sample image is normal, whether there are artifacts in the image, and whether the diagnosis area is clear by an experienced doctor as the classification gold standard; then the computer device inputs the sample image into the initial image classification network to obtain the predicted classification result, and calculates the loss between the predicted classification result and the classification gold standard, and back-propagates the loss to update the network parameters (weights) of the initial image classification network, and obtains the image classification model with training convergence through continuous iteration.

[0090] S302, based on the image classification result, calculate the quantitative result of the X-ray knee joint image under the second quality index; the second quality index is used to indicate whether the X-ray knee joint image is clear.

[0091] Specifically, based on the above-mentioned image classification results, the computer device can calculate the quantitative results of the X-ray knee joint image under the second quality index according to each classification result and the corresponding score threshold, and the second quality index is used to characterize whether the X-ray knee joint image is clear.

[0092] Optionally, the second quality index may include at least one of an image density index, an artifact index, and an image clarity index. The computer device may determine the quantification result under the image density index based on the grayscale density classification result and the preset first threshold value (such as 85 points). If the grayscale density classification result is greater than the first threshold value, it is considered that the grayscale density of the image meets the specification, otherwise it does not meet the specification. Based on the artifact classification result and the preset second threshold value (such as 60 points), if the artifact classification result is greater than the second threshold value, it is considered that there are artifacts in the image, otherwise there are no artifacts, that is, the quantification result under the artifact index is determined. Based on the clarity classification result and the preset third threshold value (such as 80 points), if the clarity classification result is greater than the third threshold value, it is considered that the image diagnosis area is clear, otherwise it is not clear, that is, the quantification result under the image clarity index is determined. It should be noted that the quantification results under the image density index, artifact index, and image clarity index in the second quality index can be calculated according to actual needs. At least one of them; in addition, for the anteroposterior knee joint image and the lateral knee joint image, the calculation method is similar.

[0093] In the quality control method for medical images provided in this embodiment, the computer device can also input the X-ray knee joint image into the image classification model to obtain an image classification result, which includes at least one of the grayscale density classification result, the artifact classification result and the clarity classification result; then based on the image classification result, the quantification result of the X-ray knee joint image under the second quality index is calculated, and the second quality index is used to characterize whether the X-ray knee joint image is clear. In this method, in addition to quantifying whether the shooting position of each tissue in the image is standardized, the image clarity is also quantified, so that the quality control angle of the medical image is more comprehensive, and it is automatically processed by the computer device without human participation, which also improves the efficiency of the quality control process.

[0094] In one embodiment, in addition to image segmentation and image classification of the X-ray knee joint image, the computer device can also perform image detection on it. Figure 5 As shown, the above method also includes:

[0095] S401, inputting the X-ray knee joint image into the image detection model to obtain the image detection result; the image detection result includes the foreign body detection result and / or the image mark detection result.

[0096] Specifically, the computer device can input the above-mentioned X-ray knee joint image into the image detection model. Through the image detection model, the non-knee joint tissue part can be detected and marked with a rectangular frame; the non-knee joint tissue part can be a foreign body (such as iatrogenic foreign body and non-iatrogenic foreign body), or a tag (including L left knee and R right knee) used to represent the left knee or right knee in the X-ray knee joint image. Optionally, the image detection model can output the iatrogenic foreign body score, the non-iatrogenic foreign body score, the left marked box score and the right marked box score, and the result corresponding to the highest score value is used as the image detection result.

[0097] Optionally, the above-mentioned image detection model can be a neural network model, such as a convolutional neural network, a recurrent neural network, etc.; optionally, the image detection model includes but is not limited to RetinaNet, Faster-RCNN, and YOLO-v3 models. Among them, the training method of the image detection model may include: obtaining a large number of sample images (the acquisition method can refer to the description of the above embodiment), and then annotating the location and type of the foreign body frame of each sample image, the location of the left and right mark frames, and the classification of the left and right marks as the detection gold standard; after the computer device inputs the sample image into the initial image detection model, the predicted detection result can be obtained by calculating the loss between the predicted detection result and the detection gold standard; the loss is back-propagated to update the network parameters (weights) of the initial image detection model, and after continuous iteration, a trained converged image detection model can be obtained.

[0098] S402, based on the image detection result, calculate the quantitative result of the X-ray knee joint image under the third quality index; the third quality index is used to characterize the type of non-tissue part in the X-ray knee joint image.

[0099] Specifically, based on the above-mentioned image detection results, the computer device can calculate the quantitative results of the X-ray knee joint image under the third quality index. Optionally, the third quality index includes a foreign body index and / or an image marking index. The computer device determines whether it is an iatrogenic foreign body based on the foreign body detection result. If the foreign body detection result is greater than a preset fourth threshold (such as 90 points), it is considered to contain an iatrogenic foreign body, and it can be considered that there is a foreign body on the image, that is, the quantitative result under the foreign body index is determined. And / or, based on the image marking detection result and the image tag (tag) of the X-ray knee joint image, it is determined whether the two correspond. If they do not correspond, it is considered that the left and right markings of the image are wrong, otherwise they are correct, that is, the quantitative result under the image marking index is determined.

[0100] It should be noted that for the above-mentioned image segmentation results, image classification results, image detection results and corresponding quantification results, the computer equipment can perform a comprehensive analysis to obtain the final quantification result of the X-ray knee joint image, such as weighted summation of each quantification result, and store each quantification result and the final quantification result together with the X-ray knee joint image into a compressed file for preservation.

[0101] In the quality control method for medical images provided in this embodiment, the computer device can also input the X-ray knee joint image into the image detection model to obtain the image detection result, which includes the foreign body detection result and / or the image marker detection result; then based on the image detection result, the quantification result of the X-ray knee joint image under the third quality index is calculated, and the third quality index is used to characterize the type of non-tissue part in the X-ray knee joint image. In this method, in addition to quantifying whether the shooting position of each tissue in the image is standardized and quantifying the clarity of the image, the type of non-tissue part in the image is also quantified, so that the quality control angle of the medical image is further comprehensive, and it is automatically processed by the computer device without human participation, which also improves the efficiency of the quality control process.

[0102] In one embodiment, the computer device can also collect statistics on the above quantification results regularly or in real time to evaluate and control the shooting quality of the corresponding doctors and improve the shooting level of the doctors. Optionally, the above method can also include:

[0103] The quantification result is sent to the client of the corresponding user in real time through the network transmission protocol, so that the user can choose whether to re-collect the X-ray knee joint image according to the quantification result;

[0104] And / or, storing the X-ray knee joint images and quantification results in a database; performing statistics on the X-ray knee joint images and quantification results in the database at preset time intervals to obtain statistical results; the statistical results represent the quality changes of the X-ray knee joint images collected by the user in the current period; and displaying the statistical results to the user's client through a web page.

[0105] Specifically, the computer device can be in a "standby state" and continuously monitor whether there is data transmitted from the shooting device to the PACS system or quality control server through the dcmtk medical image transmission protocol; after the doctor takes the X-ray knee joint image, he clicks to send it to the quality control server. The quality control server analyzes whether the received data is correct data (such as whether the tag label is complete, etc.) and puts it into the calculation queue. The computer device monitors whether there is existing data in the calculation queue, and executes the above-mentioned medical image quality control method on the existing data; the obtained quantification results are packaged and stored, and the quantification results are transmitted to the corresponding doctor's computer through network matching analysis and socket network transmission technology; the front-end software on the doctor's computer continuously monitors whether there is a quantification result package input, and decompresses and displays the quantification result package, so as to remind the doctor to retake the images that do not meet the standards.

[0106] Alternatively, the computer device is in a timed pull state, and the X-ray knee joint images are pulled from the hospital's PACS system regularly through the PACS medical image transmission protocol; after a certain number of images are pulled regularly, they are stored in the calculation queue, and the computer device monitors whether there is stock data in the calculation queue, and performs the above-mentioned medical image quality control method on the stock data; the obtained quantification results are packaged and stored, and the corresponding quantification results, quality indicators and necessary shooting information data are stored and managed through the MySQL database technology; the web background software matching the database set in the computer device can retrieve the data in the database, and the X-ray knee joint images and quantification results in the database are counted at a preset time interval (such as one week) as a cycle, and the statistical results are obtained, and the data and charts are displayed in the form of web pages. In addition, internal hospital managers can also access the specified website through different hardware and software, and then view the statistics of the quality results of the corresponding doctors' filming in any time period.

[0107] In order to better understand the processing of the above-mentioned medical image quality control method, the overall process of the method is described again below. Figure 6 As shown, the method includes:

[0108] S501, obtaining an X-ray knee joint image; the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image;

[0109] S502, inputting the X-ray knee joint image into an image segmentation model to obtain a frontal image segmentation result and / or a lateral image segmentation result;

[0110] S503, calculating the quantitative result of the anteroposterior knee joint image under the first sub-index based on the anteroposterior image segmentation result;

[0111] S504, calculating the quantitative result of the lateral knee joint image under the second sub-index based on the lateral image segmentation result;

[0112] S505, inputting the X-ray knee joint image into the image classification model to obtain an image classification result;

[0113] S506, calculating a quantitative result of the X-ray knee joint image under a second quality index based on the image classification result;

[0114] S507, inputting the X-ray knee joint image into the image detection model to obtain an image detection result;

[0115] S508, calculating a quantitative result of the X-ray knee joint image under a third quality index based on the image detection result;

[0116] S509, sending the quantification results to the corresponding user's client in real time through the network transmission protocol, so that the user can choose whether to re-collect the X-ray knee joint image according to the quantification results; and / or, at a preset time interval, performing statistics on the X-ray knee joint images and quantification results in the database to obtain statistical results, and displaying the statistical results to the user's client through a web page.

[0117] The implementation process of each step can refer to the description of the above embodiment. The implementation principle and technical effect are similar and will not be repeated here.

[0118] It should be understood that although Figure 2-Figure 6 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2-Figure 6 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0119] In one embodiment, Figure 7 As shown, a quality control device for medical images is provided, comprising: an acquisition module 11, an image segmentation module 12 and a quantification module 13.

[0120] Specifically, the acquisition module 11 is used to acquire an X-ray knee joint image, wherein the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image;

[0121] An image segmentation module 12 is used to input the X-ray knee joint image into the image segmentation model to obtain an image segmentation result; the image segmentation result includes an anteroposterior image segmentation result and / or a lateral image segmentation result, and the image segmentation result includes at least one of a femur mask, a fibula mask, a tibia mask, a patella mask, and a knee joint space mask;

[0122] The quantification module 13 is used to calculate the quantification result of the X-ray knee joint image under the first quality index based on the image segmentation result; the first quality index is used to indicate whether the shooting position of each tissue in the X-ray knee joint image is standard.

[0123] The medical image quality control device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

[0124] In one embodiment, the first quality indicator includes a first sub-indicator and / or a second sub-indicator; the quantification module 13 is specifically used to calculate the quantification result of the AP knee joint image under the first sub-indicator based on the AP image segmentation result; the first sub-indicator includes a joint long axis position indicator, a femoral intercondylar fossa position indicator, a fibula and tibia overlap indicator, a patella position indicator and at least one of the knee joint space size indicator; and / or, based on the lateral image segmentation result, calculate the quantification result of the lateral knee joint image under the second sub-indicator; the second sub-indicator includes at least one of a tibia and fibula long axis position indicator, a patellar lower edge position indicator, a femur and patella overlap indicator, a femur and fibula overlap indicator and a femur and fibula angle indicator.

[0125] In one embodiment, the quantification module 13 is specifically used to perform at least one of the following steps: determining the quantification result under the joint long axis position index based on the long axis direction of the femoral mask in the AP image segmentation result and the main axis direction of the AP knee joint image; determining the quantification result under the femoral intercondylar fossa position index based on the center point position of the lowest edge of the femoral mask in the AP image segmentation result; determining the quantification result under the fibula and tibia overlap index based on the overlapping area of ​​the fibula mask and the tibia mask in the AP image segmentation result; determining the quantification result under the patella position index based on the left and right boundaries of the patella mask and the left and right boundaries of the femur mask in the AP image segmentation result; determining the quantification result under the knee joint gap size index based on the left and right side gaps of the knee joint gap mask in the AP image segmentation result.

[0126] In one embodiment, the quantification module 13 is specifically used to perform at least one of the following steps: based on the long axis direction of the tibia mask and the fibula mask in the lateral image segmentation result and the main axis direction of the lateral knee joint image, determine the quantification result under the tibia and fibula long axis position index; based on the center point position of the lowest edge of the patella mask in the lateral image segmentation result, determine the quantification result under the patellar lower edge position index; based on the overlapping area of ​​the femur mask and the patella mask in the lateral image segmentation result, determine the quantification result under the femur and patella overlap index; based on the overlapping area of ​​the femur mask and the fibula mask in the lateral image segmentation result, determine the quantification result under the femur and fibula overlap index; based on the long axis direction of the femur mask and the long axis direction of the fibula mask in the lateral image segmentation result, determine the quantification result under the femur and fibula angle index.

[0127] In one embodiment, the above-mentioned device also includes an image classification module, which is used to input the X-ray knee joint image into an image classification model to obtain an image classification result; the image classification result includes at least one of a grayscale density classification result, an artifact classification result and a clarity classification result; the quantization module 13 is also used to calculate the quantization result of the X-ray knee joint image under a second quality index based on the image classification result; the second quality index is used to characterize whether the X-ray knee joint image is clear.

[0128] In one embodiment, the second quality indicator includes at least one of an image density indicator, an artifact indicator and an image clarity indicator; the quantization module 13 is specifically used to perform at least one of the following steps: based on the grayscale density classification result and a preset first threshold, determine the quantization result under the image density indicator; based on the artifact classification result and a preset second threshold, determine the quantization result under the artifact indicator; based on the clarity classification result and a preset third threshold, determine the quantization result under the image clarity indicator.

[0129] In one embodiment, the above-mentioned device also includes an image detection module, which is used to input the X-ray knee joint image into an image detection model to obtain an image detection result; the image detection result includes a foreign body detection result and / or an image marker detection result; the quantification module 13 is also used to calculate the quantification result of the X-ray knee joint image under a third quality indicator based on the image detection result; the third quality indicator is used to characterize the type of non-tissue part in the X-ray knee joint image.

[0130] In one embodiment, the third quality indicator includes a foreign body indicator and / or an image marking indicator; the quantification module 13 is specifically used to determine the quantification result under the foreign body indicator based on the foreign body detection result and a preset fourth threshold; and / or, based on the image marking detection result and the image label of the X-ray knee joint image, determine the quantification result under the image marking indicator.

[0131] In one embodiment, the above-mentioned device also includes an interactive module, which is used to send the quantification result to the corresponding user's client in real time through a network transmission protocol, so that the user can choose whether to re-collect the X-ray knee joint image according to the quantification result; and / or, store the X-ray knee joint image and the quantification result in a database; perform statistics on the X-ray knee joint images and the quantification results in the database at a preset time interval to obtain statistical results; the statistical results represent the quality changes of the X-ray knee joint images collected by the user in the current period; and display the statistical results to the user's client through a web page.

[0132] For the specific definition of the quality control device for medical images, please refer to the definition of the quality control method for medical images in the above text, which will not be repeated here. Each module in the above-mentioned quality control device for medical images can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0133] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 1 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a quality control method for medical images is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0134] Those skilled in the art will understand that Figure 1 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0135] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0136] Acquire an X-ray knee joint image, wherein the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image;

[0137] Inputting the X-ray knee joint image into the image segmentation model to obtain an image segmentation result; the image segmentation result includes an anteroposterior image segmentation result and / or a lateral image segmentation result, and the image segmentation result includes at least one of a femur mask, a fibula mask, a tibia mask, a patella mask, and a knee joint space mask;

[0138] Based on the image segmentation result, a quantitative result of the X-ray knee joint image under a first quality index is calculated; the first quality index is used to characterize whether the shooting position of each tissue in the X-ray knee joint image is standardized.

[0139] The implementation principle and technical effects of the computer device provided in this embodiment are similar to those of the above method embodiments, and will not be repeated here.

[0140] In one embodiment, the first quality indicator includes a first sub-indicator and / or a second sub-indicator; when the processor executes the computer program, the following steps are further implemented:

[0141] Based on the segmentation result of the anteroposterior image, a quantitative result of the anteroposterior knee joint image under the first sub-index is calculated; the first sub-index includes at least one of a joint long axis position index, a femoral intercondylar fossa position index, a fibula and tibia overlap index, a patella position index, and a knee joint space size index;

[0142] and / or,

[0143] Based on the lateral image segmentation result, the quantitative result of the lateral knee joint image under the second sub-indicator is calculated; the second sub-indicator includes at least one of the tibia and fibula long axis position index, the patellar lower edge position index, the femur and patella overlap index, the femur and fibula overlap index and the femur and fibula angle index.

[0144] In one embodiment, when the processor executes the computer program, it also implements at least one of the following steps:

[0145] Based on the long axis direction of the femoral mask in the segmentation result of the anteroposterior image and the main axis direction of the anteroposterior knee joint image, the quantitative result under the long axis position index of the joint is determined;

[0146] Based on the center point position of the lowest edge of the femoral mask in the segmentation result of the anteroposterior image, the quantitative result under the femoral intercondylar fossa position index is determined;

[0147] Based on the overlapping area of ​​the fibula mask and the tibia mask in the segmentation result of the orthotopic image, the quantitative result under the overlap index of the fibula and the tibia is determined;

[0148] Based on the left and right boundaries of the patella mask and the left and right boundaries of the femur mask in the segmentation result of the frontal image, the quantitative result under the patella position index is determined;

[0149] Based on the left and right gaps of the knee joint gap mask in the segmentation results of the anteroposterior image, the quantitative results under the knee joint gap size index are determined.

[0150] In one embodiment, when the processor executes the computer program, it also implements at least one of the following steps:

[0151] Based on the long axis directions of the tibia mask and the fibula mask in the lateral image segmentation result and the main axis direction of the lateral knee joint image, the quantitative results under the long axis position index of the tibia and fibula are determined;

[0152] Based on the center point position of the lowest edge of the patellar mask in the lateral image segmentation result, the quantitative result under the patellar lower edge position index is determined;

[0153] Based on the overlapping area of ​​the femur mask and the patella mask in the lateral image segmentation result, the quantitative result under the femur and patella overlap index is determined;

[0154] Based on the overlapping area of ​​the femur mask and the fibula mask in the lateral image segmentation result, the quantitative result under the femur and fibula overlap index is determined;

[0155] Based on the long axis direction of the femur mask and the long axis direction of the fibula mask in the lateral image segmentation result, the quantitative result under the femur-fibula angle index is determined.

[0156] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0157] Inputting the X-ray knee joint image into the image classification model to obtain an image classification result; the image classification result includes at least one of a grayscale density classification result, an artifact classification result, and a clarity classification result;

[0158] Based on the image classification result, a quantitative result of the X-ray knee joint image under a second quality index is calculated; the second quality index is used to characterize whether the X-ray knee joint image is clear.

[0159] In one embodiment, the second quality indicator includes at least one of an image density indicator, an artifact indicator, and an image clarity indicator; and when the processor executes the computer program, at least one of the following steps is further implemented:

[0160] Determine a quantization result under the image density index based on the grayscale density classification result and a preset first threshold value;

[0161] Determine a quantification result under the artifact index based on the artifact classification result and a preset second threshold;

[0162] Based on the clarity classification result and a preset third threshold, a quantization result under the image clarity index is determined.

[0163] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0164] Inputting the X-ray knee joint image into the image detection model to obtain an image detection result; the image detection result includes a foreign body detection result and / or an image marker detection result;

[0165] Based on the image detection result, the quantitative result of the X-ray knee joint image under the third quality index is calculated; the third quality index is used to characterize the type of non-tissue part in the X-ray knee joint image.

[0166] In one embodiment, the third quality indicator includes a foreign matter indicator and / or an image marking indicator; when the processor executes the computer program, the following steps are also implemented:

[0167] Based on the foreign matter detection result and a preset fourth threshold, determining a quantitative result under the foreign matter index;

[0168] and / or,

[0169] Based on the image marker detection results and the image labels of the X-ray knee joint images, the quantitative results under the image marker indicators are determined.

[0170] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0171] The quantification result is sent to the client of the corresponding user in real time through the network transmission protocol, so that the user can choose whether to re-collect the X-ray knee joint image according to the quantification result;

[0172] and / or,

[0173] Store X-ray knee joint images and quantitative results in a database;

[0174] Taking a preset time interval as a period, statistics are performed on the X-ray knee joint images and quantification results in the database to obtain statistical results; the statistical results represent the quality changes of the X-ray knee joint images collected by the user in the current period;

[0175] The statistical results are displayed to the user's client via a web page.

[0176] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0177] Acquire an X-ray knee joint image, wherein the X-ray knee joint image includes an anteroposterior knee joint image and / or a lateral knee joint image;

[0178] Inputting the X-ray knee joint image into the image segmentation model to obtain an image segmentation result; the image segmentation result includes an anteroposterior image segmentation result and / or a lateral image segmentation result, and the image segmentation result includes at least one of a femur mask, a fibula mask, a tibia mask, a patella mask, and a knee joint space mask;

[0179] Based on the image segmentation result, a quantitative result of the X-ray knee joint image under a first quality index is calculated; the first quality index is used to characterize whether the shooting position of each tissue in the X-ray knee joint image is standardized.

[0180] The computer-readable storage medium provided in this embodiment has similar implementation principles and technical effects to those of the above method embodiments, and will not be described in detail here.

[0181] In one embodiment, the first quality indicator includes a first sub-indicator and / or a second sub-indicator; when the computer program is executed by a processor, the following steps are further implemented:

[0182] Based on the segmentation result of the anteroposterior image, a quantitative result of the anteroposterior knee joint image under the first sub-index is calculated; the first sub-index includes at least one of a joint long axis position index, a femoral intercondylar fossa position index, a fibula and tibia overlap index, a patella position index, and a knee joint space size index;

[0183] and / or,

[0184] Based on the lateral image segmentation result, the quantitative result of the lateral knee joint image under the second sub-indicator is calculated; the second sub-indicator includes at least one of the tibia and fibula long axis position index, the patellar lower edge position index, the femur and patella overlap index, the femur and fibula overlap index and the femur and fibula angle index.

[0185] In one embodiment, when the computer program is executed by a processor, at least one of the following steps is also implemented:

[0186] Based on the long axis direction of the femoral mask in the segmentation result of the anteroposterior image and the main axis direction of the anteroposterior knee joint image, the quantitative result under the long axis position index of the joint is determined;

[0187] Based on the center point position of the lowest edge of the femoral mask in the segmentation result of the anteroposterior image, the quantitative result under the femoral intercondylar fossa position index is determined;

[0188] Based on the overlapping area of ​​the fibula mask and the tibia mask in the segmentation result of the orthotopic image, the quantitative result under the overlap index of the fibula and the tibia is determined;

[0189] Based on the left and right boundaries of the patella mask and the left and right boundaries of the femur mask in the segmentation result of the frontal image, the quantitative result under the patella position index is determined;

[0190] Based on the left and right gaps of the knee joint gap mask in the segmentation results of the anteroposterior image, the quantitative results under the knee joint gap size index are determined.

[0191] In one embodiment, when the computer program is executed by a processor, at least one of the following steps is also implemented:

[0192] Based on the long axis directions of the tibia mask and the fibula mask in the lateral image segmentation result and the main axis direction of the lateral knee joint image, the quantitative results under the long axis position index of the tibia and fibula are determined;

[0193] Based on the center point position of the lowest edge of the patellar mask in the lateral image segmentation result, the quantitative result under the patellar lower edge position index is determined;

[0194] Based on the overlapping area of ​​the femur mask and the patella mask in the lateral image segmentation result, the quantitative result under the femur and patella overlap index is determined;

[0195] Based on the overlapping area of ​​the femur mask and the fibula mask in the lateral image segmentation result, the quantitative result under the femur and fibula overlap index is determined;

[0196] Based on the long axis direction of the femur mask and the long axis direction of the fibula mask in the lateral image segmentation result, the quantitative result under the femur-fibula angle index is determined.

[0197] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0198] Inputting the X-ray knee joint image into the image classification model to obtain an image classification result; the image classification result includes at least one of a grayscale density classification result, an artifact classification result, and a clarity classification result;

[0199] Based on the image classification result, a quantitative result of the X-ray knee joint image under a second quality index is calculated; the second quality index is used to characterize whether the X-ray knee joint image is clear.

[0200] In one embodiment, the second quality indicator includes at least one of an image density indicator, an artifact indicator, and an image clarity indicator; and when the computer program is executed by the processor, at least one of the following steps is further implemented:

[0201] Determine a quantization result under the image density index based on the grayscale density classification result and a preset first threshold value;

[0202] Determine a quantification result under the artifact index based on the artifact classification result and a preset second threshold;

[0203] Based on the clarity classification result and a preset third threshold, a quantization result under the image clarity index is determined.

[0204] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0205] Inputting the X-ray knee joint image into the image detection model to obtain an image detection result; the image detection result includes a foreign body detection result and / or an image marker detection result;

[0206] Based on the image detection result, the quantitative result of the X-ray knee joint image under the third quality index is calculated; the third quality index is used to characterize the type of non-tissue part in the X-ray knee joint image.

[0207] In one embodiment, the third quality indicator includes a foreign matter indicator and / or an image marking indicator; when the computer program is executed by the processor, the following steps are also implemented:

[0208] Based on the foreign matter detection result and a preset fourth threshold, determining a quantitative result under the foreign matter index;

[0209] and / or,

[0210] Based on the image marker detection results and the image labels of the X-ray knee joint images, the quantitative results under the image marker indicators are determined.

[0211] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0212] The quantification results are sent to the client of the corresponding user in real time through the network transmission protocol, so that the user can choose whether to re-collect the X-ray knee joint image according to the quantification results;

[0213] and / or,

[0214] Store X-ray knee joint images and quantitative results in a database;

[0215] Taking a preset time interval as a period, statistics are performed on the X-ray knee joint images and quantification results in the database to obtain statistical results; the statistical results represent the quality changes of the X-ray knee joint images collected by the user in the current period;

[0216] The statistical results are displayed to the user's client via a web page.

[0217] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0218] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0219] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A quality control method for medical images, characterized in that: The method comprises: Acquire an X-ray knee joint image, wherein the X-ray knee joint image includes an anteroposterior knee joint image and a lateral knee joint image; Inputting the X-ray knee joint image into an image segmentation model to obtain an image segmentation result; the image segmentation result includes a frontal image segmentation result and a lateral image segmentation result, and the image segmentation result includes a femur mask, a fibula mask, a tibia mask, a patella mask and a knee joint space mask; Based on the image segmentation result, a quantitative result of the X-ray knee joint image under a first quality index is calculated; the first quality index is used to characterize whether the shooting position of each tissue in the X-ray knee joint image is standardized; the first quality index includes a first sub-indicator and a second sub-indicator; the first sub-indicator includes at least one of a joint long axis position index, a femoral intercondylar fossa position index and a knee joint space size index; the second sub-indicator includes a tibia and fibula long axis position index, a patellar lower edge position index, and a femur and patella overlap index; Wherein, calculating the quantitative result of the anteroposterior knee joint image under the first sub-index based on the segmentation result of the anteroposterior image comprises at least one of the following steps: Determine the quantitative result under the joint long axis position index based on the long axis direction of the femur mask in the orthotopic image segmentation result and the main axis direction of the orthotopic knee joint image; Determine the quantitative result under the femoral intercondylar fossa position index based on the center point position of the lower edge of the femoral mask in the orthotopic image segmentation result; Determine the quantification result under the knee joint gap size index based on the left and right gaps of the knee joint gap mask in the anteroposterior image segmentation result; Wherein, calculating the quantitative result of the lateral knee joint image under the second sub-index based on the lateral image segmentation result comprises the following steps: Determine the quantitative result under the patellar lower edge position index based on the center point position of the lower edge of the patella mask in the lateral image segmentation result; Based on the overlapping area of ​​the femur mask and the patella mask in the lateral image segmentation result, a quantification result under the femur and patella overlap index is determined.

2. The method according to claim 1, characterized in that The step of calculating the quantitative result of the lateral knee joint image under the second sub-index based on the lateral image segmentation result comprises at least one of the following steps: Based on the long axis directions of the tibia mask and the fibula mask in the lateral image segmentation result and the main axis direction of the lateral knee joint image, the quantification result under the tibia and fibula long axis position index is determined.

3. The method according to claim 1, characterized in that The evaluation index of the X-ray knee joint image also includes a second quality index, which includes at least one of an image density index, an artifact index, and an image clarity index; the image classification result of the X-ray knee joint image includes at least one of a grayscale density classification result, an artifact classification result, and a clarity classification result; the method also includes at least one of the following steps: Based on the grayscale density classification result and a preset first threshold, determining a quantization result under the image density index; Based on the artifact classification result and a preset second threshold, determining a quantization result under the artifact indicator; Based on the clarity classification result and a preset third threshold, a quantization result under the image clarity index is determined.

4. The method according to any one of claims 1 to 2, characterized in that: The method further comprises: Inputting the X-ray knee joint image into an image detection model to obtain an image detection result; the image detection result includes a foreign body detection result and / or an image mark detection result; Based on the image detection result, a quantitative result of the X-ray knee joint image under a third quality index is calculated; the third quality index is used to characterize the type of non-tissue part in the X-ray knee joint image.

5. The method according to claim 4, characterized in that The third quality index includes a foreign body index and / or an image marking index; and the step of calculating the quantitative result of the X-ray knee joint image under the third quality index based on the image detection result includes: Based on the foreign matter detection result and a preset fourth threshold, determining a quantification result under the foreign matter indicator; and / or, Based on the image tag detection result and the image tag of the X-ray knee joint image, a quantification result under the image tag index is determined.

6. The method according to claim 1, characterized in that The method further comprises: The quantification result is sent to the client of the corresponding user in real time through a network transmission protocol, so that the user can choose whether to re-collect the X-ray knee joint image according to the quantification result; and / or, storing the X-ray knee joint image and the quantification result in a database; Taking a preset time interval as a period, statistics are performed on the X-ray knee joint images and quantification results in the database to obtain statistical results; the statistical results represent the quality changes of the X-ray knee joint images collected by the user in the current period; The statistical results are displayed to the user's client via a web page.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images