A DR imaging method and a DR imaging device

Abstract

The present application provides a DR imaging method and a DR imaging device, which obtain a medical grayscale image, deeply reconstruct the medical grayscale image, obtain a display depth map, obtain a pathological decision value, make a decision cut on the display depth map according to the pathological decision value, obtain a pathological feature image set, perform imaging dimension conversion on the pathological feature image set, obtain a medical model recognition perception matrix, determine a medical perception factor according to the medical model recognition perception matrix, and then obtain a feature accumulation operator sequence, perform accumulation operator truncation estimation according to the feature accumulation operator sequence, obtain a medical truncation value, perform perception screening on the pathological feature image set according to the medical truncation value and the medical perception factor, determine a medical abnormality site information set, generate a medical imaging marking map according to the medical abnormality site information set, and improve the pathological differentiation of medical imaging of DR imaging by marking the medical abnormality sites.

Description

Technical Field

[0001] The present application relates to the field of imaging technology, and more specifically, to a DR imaging method and a DR imaging device. Background Art

[0002] In recent years, imaging methods have developed rapidly in various fields of life, such as optics, computer science, medicine, etc., especially in the field of medicine. By combining different image detection or image recognition technologies, it can be used for the inspection of internal organs such as viscera, monitoring the development of pregnant women and fetuses, etc., bringing many beneficial effects to people's lives and work.

[0003] With the development of the times, people's pace of life has accelerated. Long-term irregular diet and work and rest have caused people to suffer from various diseases. They usually go to the hospital to take B-ultrasound pictures, CT pictures and other medical imaging pictures to find out the cause of the disease. However, medical imaging pictures contain a lot of useless textures, and the pathological differentiation of the images is very low. If the clinician is inexperienced, the low pathological differentiation will make it difficult for the clinician to accurately identify the lesions in the image, and the diagnostic efficiency will be greatly reduced. Therefore, how to improve the pathological differentiation of medical imaging pictures is an urgent problem that the industry needs to solve. Summary of the invention

[0004] The present application provides a DR imaging method and a DR imaging device, which can improve the pathological differentiation of DR imaging.

[0005] In a first aspect, the present application provides a DR imaging method, comprising the following steps:

[0006] Obtain the patient's scanned image and filter it to obtain a medical grayscale image;

[0007] Performing depth reconstruction on the medical grayscale image to obtain a display depth map, obtaining a pathology decision value, and performing decision segmentation on the display depth map according to the pathology decision value to obtain a pathology feature image set;

[0008] Performing imaging dimension conversion on the pathological feature image set to obtain a medical model recognition perception matrix, and determining a medical perception factor according to the medical model recognition perception matrix;

[0009] Determine the feature accumulation operators corresponding to the pathological feature region images in the pathological feature image set, obtain a feature accumulation operator sequence, perform accumulation operator truncation estimation according to the feature accumulation operator sequence, and obtain a medical truncation value;

[0010] Performing perception screening on the pathological feature image set according to the medical cutoff value and the medical perception factor to determine a medical abnormal site information set;

[0011] A medical imaging marker map is generated according to the medical abnormality site information set.

[0012] In some embodiments, performing depth reconstruction on the medical grayscale image to obtain a display depth image specifically includes:

[0013] Get medical grayscale images;

[0014] Performing pixel reorganization on the medical grayscale image to determine a medical pixel image;

[0015] The medical pixel image is subjected to depth frequency decomposition to obtain a display depth image.

[0016] In some embodiments, obtaining the pathology decision value may specifically include:

[0017] determining a medical pixel attribute value according to the displayed depth map;

[0018] Obtain all medical high-frequency pixel components in the displayed depth map;

[0019] Obtain all medical low-frequency pixel components in the displayed depth map;

[0020] The pathology decision value is determined according to the medical high-frequency pixel component, the medical low-frequency pixel component and the medical pixel attribute value.

[0021] In some embodiments, the display depth map is segmented according to the pathology decision value to obtain a pathology feature image set, which can be achieved by the following steps:

[0022] The displayed depth map is segmented and decided according to the pathological decision value to obtain a plurality of medical edge images;

[0023] All medical edge images are binarized to obtain a set of pathological feature images.

[0024] In some embodiments, performing imaging dimension conversion on the pathological feature image set to obtain a medical model recognition perception matrix can be achieved by using the following steps:

[0025] Acquire all pathological characteristic region images in a pathological characteristic image set;

[0026] Determine the medical feature matrix from all pathological feature area maps;

[0027] Perform dimensionality reduction on the medical feature matrix to determine the medical model recognition matrix.

[0028] In some embodiments, principal component analysis is used to perform dimensionality reduction processing on the medical feature matrix to determine the medical model recognition matrix.

[0029] In some embodiments, the following steps may be used to obtain a patient scan image and filter it to obtain a medical grayscale image:

[0030] Obtain patient scan images;

[0031] Equalizing the patient scanned image to obtain a medical enhancement histogram;

[0032] The medical enhancement histogram is filtered to determine a medical grayscale image.

[0033] In a second aspect, the present application provides a DR imaging device, which includes a medical imaging image control unit, wherein the medical imaging image control unit includes:

[0034] An acquisition module is used to acquire and filter the patient's scanned image to obtain a medical grayscale image;

[0035] A processing module, configured to perform depth reconstruction on the medical grayscale image to obtain a display depth image, obtain a pathology decision value, perform decision segmentation on the display depth image according to the pathology decision value, and obtain a pathology feature image set;

[0036] The processing module is further used to perform imaging dimension conversion on the pathological feature image set to obtain a medical model recognition matrix, and determine a medical recognition factor according to the medical model recognition matrix;

[0037] The processing module is further used to determine the feature accumulation operators corresponding to the pathological feature region images in the pathological feature image set, obtain a feature accumulation operator sequence, and perform accumulation operator truncation estimation according to the feature accumulation operator sequence to obtain a medical truncation value;

[0038] The processing module is further used to perform perception screening on the pathological feature image set according to the medical cutoff value and the medical perception factor to determine the medical abnormal site information set;

[0039] An imaging module is used to generate a medical imaging marker map according to the medical abnormality site information set.

[0040] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned DR imaging method.

[0041] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned DR imaging method is implemented.

[0042] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects:

[0043] In the DR imaging method and DR imaging device provided by the present application, a patient scan image is first obtained and filtered to obtain a medical grayscale image, the medical grayscale image is deeply reconstructed to obtain a display depth map, a pathological decision value is obtained, and a decision segmentation is performed on the display depth map according to the pathological decision value to obtain a pathological feature image set. The decision segmentation can retain the edge information of objects in the image, improve the accuracy of abnormal point detection and positioning, perform imaging dimension conversion on the pathological feature image set to obtain a medical model recognition perception matrix, determine the medical perception factor according to the medical model recognition perception matrix, determine the feature accumulation operators corresponding to each pathological feature region map in the pathological feature image set, obtain a feature accumulation operator sequence, perform accumulation operator truncation estimation according to the feature accumulation operator sequence to obtain a medical truncation value, perform perception screening on the pathological feature image set according to the medical truncation value and the medical perception factor, determine a medical abnormal site information set, generate a medical imaging marking map according to the medical abnormal site information set, and improve the pathological differentiation of DR imaging by marking the medical abnormal sites. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is an exemplary flow chart of a method for implementing DR imaging shown in some embodiments of the present application;

[0045] Figure 2 is an exemplary flow chart for implementing depth reconstruction of a display depth map in some embodiments of the present application;

[0046] Figure 3 is an exemplary flow chart for implementing imaging dimension conversion in some embodiments of the present application;

[0047] Figure 4 is a schematic diagram of exemplary hardware and / or software for implementing the medical imaging control unit shown in some embodiments of the present application;

[0048] Figure 5 It is a schematic diagram of the structure of a computer device for implementing the method for DR imaging shown in some embodiments of the present application. DETAILED DESCRIPTION

[0049] The core of the present application is to obtain a patient scanned image and filter it to obtain a medical grayscale image, deeply reconstruct the medical grayscale image to obtain a display depth map, obtain a pathological decision value, and make a decision cut on the display depth map according to the pathological decision value to obtain a pathological feature image set. The decision cut can retain the edge information of objects in the image, improve the accuracy of abnormal point detection and positioning, perform imaging dimension conversion on the pathological feature image set to obtain a medical model recognition perception matrix, determine the medical perception factor according to the medical model recognition perception matrix, determine the feature accumulation operators corresponding to each pathological feature area map in the pathological feature image set, obtain a feature accumulation operator sequence, perform accumulation operator truncation estimation according to the feature accumulation operator sequence, obtain a medical truncation value, perform perception screening on the pathological feature image set according to the medical truncation value and the medical perception factor, determine a medical abnormal site information set, generate a medical imaging marking map according to the medical abnormal site information set, and improve the pathological differentiation of DR imaging by marking the medical abnormal sites.

[0050] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Figure 1 , which is an exemplary flow chart of a DR imaging method according to some embodiments of the present application, and the DR imaging method 100 mainly includes the following steps:

[0051] In step 101, a patient scan image is acquired and filtered to obtain a medical grayscale image.

[0052] In some embodiments, the patient scan image can be collected by medical detection equipment (such as X-ray machine, CT scanner, MRI, etc.) and uploaded to the medical cloud service system. The medical display obtains the patient scan image from the medical system, and the patient scan image is transmitted to the medical display in digital form. In specific implementation, after obtaining the patient's permission at the cloud service center, the patient scan image can be obtained from the medical system through application programming interface (API) data call, crawler and other data call methods.

[0053] Preferably, in some embodiments, obtaining a patient scan image and filtering it to obtain a medical grayscale image can be achieved by the following steps:

[0054] Obtain patient scan images;

[0055] Equalizing the patient scanned image to obtain a medical enhancement histogram;

[0056] The medical enhancement histogram is filtered to determine a medical grayscale image.

[0057] In specific implementation, in some embodiments, the patient scanned image is enhanced by a histogram equalization method to improve the visualization effect of the image and obtain a medically enhanced histogram. The medically enhanced histogram can be filtered by a Gaussian filter commonly used in the prior art for filtering images. For example, a weighted average of all pixel values ​​in the medically enhanced histogram is calculated, and the weighted average of all pixel values ​​is used as a Gaussian kernel. The obtained Gaussian kernel is then convolved with the medically enhanced histogram, thereby effectively removing high-frequency noise in the medical imaging image through Gaussian filtering, while retaining the edges and details of the medical imaging image, thereby determining a medical grayscale image.

[0058] It should be noted that in other embodiments, other filtering methods may also be used to process the image, for example, median filtering, mean filtering and other filtering methods. In addition, the convolution operation in the present application is to place a Gaussian kernel at each pixel position of the image, multiply each element of the Gaussian kernel by the image pixel value at the corresponding position and then add them to obtain a new pixel value. This process can be implemented through existing technologies such as OpenCV and MATLAB, and will not be repeated here.

[0059] Preferably, it should be noted that filtering the patient scanned image to determine the medical grayscale image reduces the interference of noise on abnormal point detection, thereby obtaining a smooth medical grayscale image to facilitate subsequent further processing of the image.

[0060] In step 102, the medical grayscale image is deeply reconstructed to obtain a display depth map, a pathology decision value is obtained, and the display depth map is decision-cut according to the pathology decision value to obtain a pathology feature image set.

[0061] In some embodiments, reference Figure 2 As shown, this figure is an exemplary flow chart of performing depth reconstruction on a display depth map in some embodiments of the present application, and performing depth reconstruction on the medical grayscale image to obtain a display depth map specifically includes:

[0062] In step S1021, a medical grayscale image is obtained;

[0063] In step S1022, pixel reorganization is performed on the medical grayscale image to determine a medical pixel image;

[0064] In step S1023, the medical pixel image is subjected to depth frequency decomposition to obtain a display depth map.

[0065] In some embodiments, the medical grayscale image is pixel-reorganized to determine the medical pixel map, specifically comprising: obtaining the medical pixel map by rearranging the pixels in the medical pixel map, that is, extracting all the pixels in the medical grayscale image, arranging and combining the extracted pixels, for example, symmetrically exchanging the pixels in the left half of the medical grayscale image with the pixels in the right half of the medical grayscale image to achieve pixel reorganization of the image and obtain the medical pixel map. In other embodiments, pixel reorganization can also be performed by using other rules for arranging and combining rows or columns of pixels. It should be noted that pixel reorganization is an image processing method that enhances image contrast information by recombining the pixel arrangement in the original image. In the present application, pixel reorganization of the medical grayscale image can obtain an image with more texture details.

[0066] In some embodiments, during the process of performing depth frequency decomposition on the medical pixel map to obtain a display depth map, in a specific implementation, the medical pixel map is first acquired, and the medical pixel map is decomposed by a wavelet basis function to obtain different frequency components, wherein the decomposed frequency components include multiple medical high-frequency pixel components and multiple medical low-frequency pixel components, the medical low-frequency pixel components include the structure and global features of the image, and the medical high-frequency pixel components include the details and texture information of the medical image, and then the high-frequency frequency components are selected in the frequency domain, and a high-frequency frequency domain threshold is set, and the medical high-frequency pixel components greater than the high-frequency frequency domain threshold are eliminated to reduce the interference of noise generated in the processing process on the image, and the decomposed frequency components are inversely transformed to obtain a display depth map.

[0067] It should be noted that the wavelet basis functions and inverse transforms in the present application are processed using wavelet transforms in the prior art. Wavelet transforms can decompose images into components of different frequencies, and different components correspond to details and structures at different frequencies. Other deep frequency decomposition methods may also be used in other embodiments, which are not limited here.

[0068] In some embodiments, obtaining the pathology decision value may specifically include:

[0069] determining a medical pixel attribute value according to the displayed depth map;

[0070] Obtain all medical high-frequency pixel components in the displayed depth map;

[0071] Obtain all medical low-frequency pixel components in the displayed depth map;

[0072] The pathology decision value is determined according to the medical high-frequency pixel component, the medical low-frequency pixel component and the medical pixel attribute value, wherein the pathology decision value can be determined according to the following formula:

[0073]

[0074] Wherein, y represents the weighted average value of medical high-frequency pixel components, x represents the weighted average value of medical low-frequency pixel components, μ represents the medical pixel attribute value, and exp represents the base of the natural logarithm e raised to the power of x.

[0075] The pathological decision value is a decision parameter used to divide the pixels in the medical image into two or more different regions, and is used to determine whether the pixel point in the image is a target (feature) region or a background region.

[0076] It should be noted that the medical pixel attribute value is the average value of the pixel values ​​of the entire display depth map. For example, each pixel value in the display depth map is added to obtain a total pixel value, and the total pixel value is divided by the total number of pixels to determine the pixel attribute value of the display depth map, which reflects the average level of the intensity of the acoustic wave echo in the display depth map. The medical pixel attribute value can be calculated by image processing libraries such as OpenCV and Pillow in the prior art;

[0077] In addition, it should be noted that the weighted average values ​​of the medical high-frequency pixel components and the medical low-frequency pixel components are calculated in order to simplify the calculations, using programming languages ​​such as Python and Javascript in the prior art. Other calculation methods may also be used in other embodiments, which will not be described in detail here.

[0078] Preferably, in some embodiments, the display depth map is segmented according to the pathology decision value to obtain the pathology feature image set, which can be achieved by the following steps:

[0079] The displayed depth map is segmented and decided according to the pathological decision value to obtain a plurality of medical edge images;

[0080] All medical edge images are binarized to determine the pathological feature image set.

[0081] In a specific implementation, the display depth map is decision-cut according to the pathological decision value, that is, the medical high-frequency pixel components in the display depth map are screened using the pathological decision value as a segmentation threshold, and the larger medical high-frequency pixel components are retained, thereby realizing decision analysis of the display depth map, that is, analyzing the pixel points of various texture information, shape information and other features existing in the display depth map, and then performing wavelet inverse transformation on the screened medical high-frequency pixel components, connecting all the image pixels after the inverse transformation to obtain edge contours and connected areas, and segmenting the edge contours in the display depth map, realizing decision cutting of the display depth map, and thus obtaining multiple medical edge images;

[0082] All medical edge images are binarized. According to the preset medical binarization threshold, the pixels in all medical edge images that are greater than the medical binarization threshold are marked as feature pixels, and the other pixels are marked as background. The maximum connected area composed of the feature pixels in each medical edge image is used as the pathological feature area map corresponding to the medical edge image, and then the pathological feature area maps corresponding to all medical edge images are determined, and combined to obtain a pathological feature image set.

[0083] In addition, it should be noted that a depth reconstruction method is used on the image to obtain a display depth map, which provides more comprehensive and detailed image information than a planar medical image, retains the structural information of the medical image, can better identify image edge information, and improves the accuracy of identifying abnormal points in the image.

[0084] The pathological feature area map is obtained by decision-making and screening the display depth map, thereby segmenting the image containing more local details in the display depth map from the image background. The pathological feature area map is used for targeted detection and monitoring of lesions, which can avoid the interference of the background area on the imaging results during the medical image imaging process.

[0085] In step 103, the imaging dimension of the pathological feature image set is converted to obtain a medical model recognition perception matrix, and a medical perception factor is determined according to the medical model recognition perception matrix.

[0086] Optionally, in some embodiments, reference Figure 3 As shown in FIG. 1 , this figure is an exemplary flow chart for implementing imaging dimension conversion in some embodiments of the present application. The imaging dimension conversion of the pathological feature image set to obtain the medical model recognition perception matrix can be implemented by the following steps:

[0087] In step S1031, all pathological characteristic region images in the pathological characteristic image set are obtained;

[0088] In step S1032, a medical feature matrix is ​​determined from all pathological feature region maps;

[0089] In step S1033, the medical feature matrix is ​​subjected to dimensionality reduction processing to determine a medical model recognition matrix.

[0090] In a specific implementation, each column of the matrix corresponds to a pathological feature region map, and the matrix element value in the column represents the pixel value of the characteristic pixel point in the pathological feature region map, thereby determining the medical feature matrix, performing eigenvalue decomposition on the medical feature matrix, and obtaining multiple eigenvalues ​​and eigenvectors corresponding to the eigenvalues;

[0091] According to the size of the eigenvalue, the first w largest eigenvalues ​​and the eigenvector corresponding to the maximum eigenvalue are selected to perform dimensionality reduction processing on the medical feature matrix, thereby determining the medical model recognition perception matrix, where w is the dimensionality reduction depth, which is usually calibrated to a positive integer based on the dimensionality reduction requirements. The medical model recognition perception matrix is ​​mainly used to characterize the distribution of feature pixel points in medical images.

[0092] Optionally, in some embodiments, the number of rows of the medical feature matrix is ​​determined according to the size of the largest pathological feature area map in the pathological feature image set. To facilitate subsequent dimensionality reduction of the medical feature matrix, when there are empty elements in the medical feature matrix, the empty elements are treated as zero elements and matrix dimensionality reduction is performed. It should be noted that the characteristic pixel points in the present application are pixel points of pathological features such as abnormal point structures, cystic structures, and abnormal accumulation of fluid due to the abnormal distribution of image pixels in the medical imaging image.

[0093] In addition, it should be noted that, in some embodiments, when performing dimensionality reduction processing on the medical feature matrix and determining the medical model recognition matrix, the dimensionality reduction processing can be performed using the PCA method in the prior art. In specific implementation, the eigenvectors corresponding to the first W largest eigenvalues ​​can also be used to form a medical sample matrix U w , w is the dimensionality reduction depth, and the transposed matrix of the medical sample matrix is ​​multiplied by the medical sample matrix to complete the dimensionality reduction processing of the medical feature matrix and obtain the medical modality recognition perception matrix, wherein the medical modality recognition perception matrix can be determined according to the following formula:

[0094] M=(U w ) T Q

[0095] Among them, (U w ) T represents the transposed matrix of the medical sample matrix, Q represents the medical feature matrix, M represents the medical pattern recognition matrix, and other dimensionality reduction methods may also be used in other embodiments, which are not limited here.

[0096] Preferably, in some embodiments, determining the medical perception factor according to the medical model perception matrix can be implemented by the following steps:

[0097] Determining a medical covariance matrix according to the medical model recognition matrix;

[0098] Determining a covariance calibration component according to the medical covariance matrix;

[0099] Determining the maximum eigenvalue of the medical covariance matrix;

[0100] Get medical pixel attribute values;

[0101] The medical perception factor is determined according to the covariance calibration component, the maximum eigenvalue of the medical covariance matrix and the medical pixel attribute value, wherein the medical perception factor α is determined by the following formula:

[0102]

[0103] Among them, λ max represents the maximum eigenvalue of the medical covariance matrix, β represents the covariance calibration component, and μ represents the medical pixel attribute value.

[0104] A medical covariance matrix is ​​determined according to the medical feature matrix. In some embodiments, the process of determining the medical covariance matrix according to the medical model recognition perception matrix can be performed by estimating the covariance matrix of the medical model recognition perception matrix to reduce the degree of discreteness between different feature pixel values ​​in the medical model recognition perception matrix, thereby obtaining the medical covariance matrix. It should be noted that the covariance matrix estimation is a data processing method that performs statistical analysis on a given data matrix to calculate the covariance relationship between each variable, thereby generating a covariance matrix.

[0105] It should be noted that the covariance calibration component is the first vector component of the eigenvector corresponding to the maximum eigenvalue of the medical covariance matrix, that is, the first dimensional element value of the eigenvector in the vector space. In some embodiments, the first vector component in the eigenvector can be obtained through the index in the Python library and used as a numerical value to facilitate calculation and processing.

[0106] The maximum eigenvalue of the medical covariance matrix can be determined by performing eigendecomposition on the medical covariance matrix, determining multiple eigenvalues ​​of the medical covariance matrix, and selecting the maximum eigenvalue from the multiple eigenvalues. In some embodiments, the maximum eigenvalue of the medical covariance matrix can be calculated based on NumPy in the Python library.

[0107] In addition, it should be noted that the medical perception factor is a feature recognition coefficient determined based on the maximum eigenvalue of the medical covariance matrix, the covariance calibration component and the medical pixel attribute value, and is used to identify abnormal features of pixel points in the medical imaging image based on the discrimination of each pathological feature area map.

[0108] In step 104, feature accumulation operators corresponding to the pathological feature region images in the pathological feature image set are determined to obtain a feature accumulation operator sequence, and accumulation operator truncation estimation is performed according to the feature accumulation operator sequence to obtain a medical truncation value.

[0109] It should be noted that the feature accumulation operator sequence is composed of feature accumulation operators corresponding to each pathological feature region map. In some embodiments, in the process of determining the feature accumulation operators corresponding to each pathological feature region map in the pathological feature image set and obtaining the feature accumulation operator sequence, the determination of the medical feature accumulation operator corresponding to the i-th pathological feature region map in the medical feature accumulation operator sequence can be achieved by the following steps:

[0110] Get the co-square calibration component;

[0111] Get medical pixel attribute values;

[0112] Determine the recognition coefficient of medical images;

[0113] Determine the dissimilarity value of medical images;

[0114] The medical feature accumulation operator sequence is determined according to the co-square calibration component, the medical pixel attribute value, the medical image recognition coefficient and the medical image differentiability value, wherein the i-th pathological

[0115] Medical feature accumulation operator ξ corresponding to the feature region graph i It can be determined according to the following formula:

[0116]

[0117] Among them, c represents the medical image recognition coefficient, β represents the covariance calibration component, μ represents the medical pixel attribute value, ε represents the medical image differentiability value, and d i is the pixel mean of the i-th pathological feature area map.

[0118] It should be noted that the medical image recognition coefficient in the present application is a quantitative value used to measure the number, size, distribution and other attributes of specific abnormal features in medical images. The larger the medical image recognition coefficient, the more obvious the abnormal features are, and the smaller the medical recognition coefficient, the less obvious the abnormal features are. Under normal circumstances, the medical image recognition coefficient can be obtained by calculating the machine model. The ratio between the correct recognition rate and the error recognition rate of historical image recognition is determined by the machine model as the medical image recognition coefficient. In some embodiments, the medical image recognition coefficient can be calibrated to a constant of 2.

[0119] The medical image dissimilarity value is a quantitative value used to determine whether a feature point in a pathological feature region map is an abnormal point. In specific implementation, the medical image dissimilarity value can be determined by the medical pixel density value, the medical pixel comparison value, the medical pixel attribute value and the medical abnormality weight coefficient, wherein the medical image dissimilarity value can be determined according to the following formula:

[0120]

[0121] Among them, ρ represents the medical feature pixel density value, γ represents the medical pixel difference value, represents the medical abnormality weight coefficient, and μ represents the medical pixel attribute value.

[0122] The feature accumulation operators corresponding to all the pathological feature region images in the pathological feature image set are obtained, all the feature accumulation operators are arranged in order from small to large, and a feature accumulation operator sequence is determined.

[0123] It should be noted that the medical feature pixel density value represents the density quantification value of the feature pixel points in the medical imaging image, that is, the ratio of the number of feature pixels to that of ordinary pixels in the medical imaging image; the medical pixel difference value is the average value of the difference in the number of abnormal feature points and normal pixels in the patient's scanned image obtained after machine learning; the medical abnormality weight coefficient represents the degree of influence of different abnormal features on the medical imaging image, and the proportional coefficient between the weighted average value of all abnormal feature pixels in the historical medical imaging image and the average value of all pixels in the historical medical imaging image is calculated, and the proportional coefficient is used as the medical abnormality weight coefficient; the medical feature pixel density value, the medical pixel difference value and the medical abnormality weight coefficient can all be obtained using Python, JavaScript, and other existing technologies. v a. Calculated and determined by C++ and other programming software, which will not be elaborated here.

[0124] In some embodiments, the following steps may be used to perform accumulation operator cutoff estimation on the feature accumulation operator sequence to obtain a medical cutoff value:

[0125] Get medical pixel attribute values;

[0126] Acquire the feature accumulation operators corresponding to the pathological feature region images in the feature accumulation operator sequence;

[0127] The medical cutoff value is determined according to the medical pixel attribute value and the feature accumulation operator corresponding to each pathological feature region map in the feature accumulation operator sequence, wherein the medical cutoff value η can be determined by the following formula:

[0128]

[0129] Wherein, N represents the total number of characteristic pixels in all pathological characteristic region maps, μ represents the medical pixel attribute value, A is the sequence length of the characteristic accumulation operator sequence, ξ i is the medical feature accumulation operator corresponding to the i-th pathological feature region map in the medical feature accumulation operator sequence, O i为第 The number of pixels in the i pathological feature region map.

[0130] It should be noted that the medical cutoff value is a determination coefficient for determining the degree of abnormality of pixels in a medical image, and is used to determine abnormal features of pixels in a medical image based on an average level of pixel values ​​in the pathological feature image set.

[0131] Optionally, in some embodiments, the sequence mean of the feature accumulation operator sequence may also be used as the medical cutoff value.

[0132] In step 105, the pathological feature image set is subjected to perception screening according to the medical cutoff value and the medical perception factor to determine a medical abnormal site information set.

[0133] In some embodiments, the pathological feature image set is subjected to perception screening according to the medical cutoff value and the medical perception factor, and the determination of the medical abnormal site information set can be achieved by the following steps:

[0134] Traversing and identifying the pathological characteristic region map according to the medical perception factor to determine the middle pixel point;

[0135] The medical cutoff value is used to perform abnormal discrimination on the intermediate pixel points, thereby determining a medical abnormal site information set.

[0136] First, the pixel points in the pathological feature region map are identified and judged according to the medical perception factor, and the medical perception ratio coefficient is obtained by comparing the pixel points in the pathological feature region map with the medical perception factor. In specific implementation, the ratio of the medical perception factor to the average pixel value of all the pixel points in the pathological feature region map can be used as the medical perception ratio coefficient, and then the pixel points in the pathological feature region map are traversed and identified by the medical perception ratio coefficient to determine the middle pixel point. In specific implementation, the pixel point whose ratio with the pixel value of the other pixel points in the neighborhood is higher than the medical perception ratio coefficient can be used as the middle pixel point. The middle pixel point is the pixel point with a drastically changing feature point image in the pathological feature region map, and thus contains most of the pathological features and texture features in the pathological feature region map;

[0137] Optionally, in some embodiments, in the process of using the medical cutoff value to perform abnormal discrimination on the intermediate pixel points and then determining the medical abnormality site information set, the medical cutoff value can be used to determine whether the intermediate pixel point is an abnormal point. For example, the intermediate pixel point greater than the medical cutoff value is taken as a medical abnormality point, and all medical abnormality points are combined to obtain a medical abnormality site information set. The medical abnormality point is the pixel point in the pathological feature area map that is most likely to contain pathological features. The medical abnormality point is highlighted, which can improve the doctor's efficiency in diagnosing the patient's lesions.

[0138] Preferably, in some embodiments, after determining the medical abnormality site information set, the method further includes: performing denoising on the medical image in the medical abnormality site information set to reduce possible noise or artifacts, performing connected region analysis on the denoised medical image, identifying and extracting connected regions, extracting abnormal point location information from the connected regions, such as calculating the center coordinates, bounding box or other features of the connected regions, organizing the extracted abnormal sites into a set, and using the set as the medical abnormality site information set for subsequent analysis and use.

[0139] It should be noted that the noise reduction processing in the present application adopts the deep learning method in the prior art for noise reduction. By using deep learning models such as convolutional neural networks (CNN) or generative adversarial networks (GAN) to reduce the noise of images, it has good processing capabilities for various types of noise and complex image structures. Other noise reduction methods can also be used in other embodiments, which are not limited here.

[0140] In step 106, a medical imaging marker map is generated according to the medical abnormality site information set.

[0141] It should be noted that pathological differentiation = the sum of pixel values ​​of pixels containing lesion features / (the sum of pixel values ​​of pixels containing lesion features + the sum of pixel values ​​of useless pixels). The higher the pathological differentiation, the more prominent the lesion area in the medical imaging image is. In the present application, the abnormal sites of the medical abnormal site information set in the medical imaging image are highlighted, thereby improving the pathological differentiation of the medical imaging image and increasing the efficiency of doctors in judging lesions in patients.

[0142] In some embodiments, during specific implementation, according to the medical abnormal site information set, image processing software (such as the Matplotlib library in the Python library) locates and marks abnormal pixel points on the medical imaging image. The abnormal pixel point refers to a pixel point that may have pathological characteristics such as abnormal point-like structures, cystic structures, and abnormal accumulation of fluid. The marked abnormal pixel point is used as an abnormal site. During specific implementation, the abnormal pixel point can be marked by increasing the pixel value of the abnormal pixel point, so that the abnormal site on the medical imaging image can be intuitively identified, and the marked medical imaging image is superimposed with the original medical imaging image to ensure that the abnormal site matches the medical imaging image, and then the superimposed image is visualized to ensure that the abnormal site is clearly visible in the image, and finally a medical imaging marking image is generated. In the generated medical imaging marking image, the pathological differentiation of the medical imaging image is improved, and the efficiency of doctors in judging lesions in patients is increased.

[0143] It should be noted that the visualization in this application is to adjust the image parameters such as color mapping, brightness, contrast, etc., which can be achieved through existing technologies such as image processing software, programming languages ​​or professional medical image processing, and will not be repeated here.

[0144] In addition, in another aspect of the present application, in some embodiments, the present application provides a DR imaging device, the system includes a medical imaging control unit, referring to Figure 4 , which is a schematic diagram of exemplary hardware and / or software of a medical imaging image control unit according to some embodiments of the present application, the medical imaging image control unit 200 includes: an acquisition module 201, a processing module 202 and an imaging module 203, which are described as follows:

[0145] Acquisition module 201, in this application, acquisition module 201 is mainly used to acquire and filter the patient scan image to obtain a medical grayscale image;

[0146] Processing module 202, in the present application, the processing module 202 is mainly used to perform depth reconstruction on the medical grayscale image, obtain a display depth map, obtain a pathology decision value, perform decision segmentation on the display depth map according to the pathology decision value, and obtain a pathology feature image set;

[0147] It should be noted that the processing module 202 in the present application is also used to perform imaging dimension conversion on the pathological feature image set to obtain a medical model recognition matrix, and determine a medical recognition factor according to the medical model recognition matrix;

[0148] In addition, the processing module 202 in the present application is also used to determine the feature accumulation operators corresponding to each pathological feature region map in the pathological feature image set, obtain a feature accumulation operator sequence, perform accumulation operator truncation estimation according to the feature accumulation operator sequence, and obtain a medical truncation value;

[0149] In addition, the processing module 202 in the present application is also used to perform perception screening on the pathological feature image set according to the medical cutoff value and the medical perception factor to determine the medical abnormal site information set;

[0150] Imaging module 203, in the present application, the imaging module 203 is mainly used to generate a medical imaging marker map according to the medical abnormality site information set.

[0151] In addition, the present application also provides a computer device, which includes a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned DR imaging method.

[0152] In some embodiments, reference Figure 5, which is a schematic diagram of the structure of a computer device for applying a DR imaging method according to some embodiments of the present application. The DR imaging method in the above embodiment can be Figure 5 The computer device 300 shown in the figure is implemented, and the computer device 300 includes at least one processor 301, a communication bus 302, a memory 303 and at least one communication interface 304.

[0153] The processor 301 may be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more processors for controlling the execution of the DR imaging method of the present application.

[0154] Communication bus 302 may include a pathway for transmitting information between the above-mentioned components.

[0155] The memory 303 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a disk or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 303 may exist independently and be connected to the processor 301 via the communication bus 302. The memory 303 may also be integrated with the processor 301.

[0156] The memory 303 is used to store the program code for executing the solution of the present application, and the execution is controlled by the processor 301. The processor 301 is used to execute the program code stored in the memory 303. The program code may include one or more software modules. The method of DR imaging in the above embodiment can be implemented by the processor 301 and one or more software modules in the program code in the memory 303.

[0157] The communication interface 304 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0158] In a specific implementation, as an embodiment, a computer device may include multiple processors, each of which may be a single-CPU processor or a multi-CPU processor. The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0159] The above-mentioned computer device can be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device can be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device or an embedded device. The embodiment of the present application does not limit the type of computer device.

[0160] In addition, the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned DR imaging method is implemented.

[0161] In summary, in a DR imaging method and a DR imaging device disclosed in an embodiment of the present application, a patient scan image is first obtained and filtered to obtain a medical grayscale image, the medical grayscale image is deeply reconstructed to obtain a display depth map, a pathological decision value is obtained, and a decision segmentation is performed on the display depth map according to the pathological decision value to obtain a pathological feature image set. The decision segmentation can retain the edge information of objects in the image, improve the accuracy of abnormal point detection and positioning, perform imaging dimension conversion on the pathological feature image set to obtain a medical model recognition perception matrix, determine the medical perception factor according to the medical model recognition perception matrix, determine the feature accumulation operators corresponding to each pathological feature region map in the pathological feature image set, obtain a feature accumulation operator sequence, perform accumulation operator truncation estimation according to the feature accumulation operator sequence, obtain a medical truncation value, perform perception screening on the pathological feature image set according to the medical truncation value and the medical perception factor, determine a medical abnormal site information set, and generate a medical imaging marking map according to the medical abnormal site information set. By marking the medical abnormal sites, the pathological differentiation of DR imaging is improved, and the efficiency of doctors in judging lesions in patients is improved.

[0162] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0163] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A DR imaging method, characterized in that: include: Obtain the patient's scanned image and filter it to obtain a medical grayscale image; Performing depth reconstruction on the medical grayscale image to obtain a display depth map, obtaining a pathology decision value, and performing decision segmentation on the display depth map according to the pathology decision value to obtain a pathology feature image set; Performing imaging dimension conversion on the pathological feature image set to obtain a medical model recognition perception matrix, and determining a medical perception factor according to the medical model recognition perception matrix; Determine the feature accumulation operators corresponding to the pathological feature region images in the pathological feature image set, obtain a feature accumulation operator sequence, perform accumulation operator truncation estimation according to the feature accumulation operator sequence, and obtain a medical truncation value; Performing perception screening on the pathological feature image set according to the medical cutoff value and the medical perception factor to determine a medical abnormal site information set; generating a medical imaging marker map according to the medical abnormality site information set; Obtaining pathological decision values ​​specifically includes: determining a medical pixel attribute value according to the displayed depth map; Obtain all medical high-frequency pixel components in the displayed depth map; Obtain all medical low-frequency pixel components in the displayed depth map; Determine a pathology decision value according to a medical high-frequency pixel component, a medical low-frequency pixel component and a medical pixel attribute value; The displayed depth map is determined and segmented according to the pathological decision value to obtain a pathological feature image set, which specifically includes: The displayed depth map is segmented and decided according to the pathological decision value to obtain a plurality of medical edge images; Binarize all medical edge images to obtain a set of pathological feature images; The imaging dimension conversion of the pathological feature image set is performed to obtain a medical model recognition perception matrix, which specifically includes: Acquire all pathological characteristic region images in a pathological characteristic image set; Determine the medical feature matrix from all pathological feature area maps; Perform dimensionality reduction processing on the medical feature matrix to determine the medical model recognition matrix; The principal component analysis method is used to reduce the dimension of the medical feature matrix and determine the medical model recognition matrix; Among them, the pathological decision value can be determined according to the following formula: in, Represents the weighted average of medical high-frequency pixel components, Represents the weighted average of medical low-frequency pixel components, Represents the medical pixel attribute value, Represents the base of natural logarithms of Power; The medical model recognition matrix can be determined according to the following formula: in, represents the transposed matrix of the medical sample matrix, represents the medical feature matrix, represents the medical model recognition matrix; Among them, medical awareness factor Determined by the following formula: in, represents the maximum eigenvalue of the medical covariance matrix, represents the co-square calibration component, Represents medical pixel attribute value; Among them, the first Medical feature accumulation operator corresponding to the pathological feature region map It can be determined according to the following formula: in, represents the medical image recognition coefficient, represents the co-square calibration component, Represents the medical pixel attribute value, represents the medical image heterogeneity value, For the The pixel mean of the pathological feature area map; Among them, the medical cutoff value It can be determined by the following formula: in, Represents the total number of characteristic pixels in all pathological characteristic area maps, represents the medical pixel attribute value, A is the sequence length of the feature accumulation operator sequence, is the first in the medical feature accumulation operator sequence The medical feature accumulation operator corresponding to the pathological feature region map, For the The number of pixels in the pathological feature area map.

2. The method according to claim 1, characterized in that Performing depth reconstruction on the medical grayscale image to obtain a display depth image specifically includes: Get medical grayscale images; Performing pixel reorganization on the medical grayscale image to determine a medical pixel image; The medical pixel image is subjected to depth frequency decomposition to obtain a display depth image.

3. The method according to claim 1, characterized in that Obtain the patient's scanned image and filter it to obtain a medical grayscale image, which specifically includes: Obtain patient scan images; Equalizing the patient scanned image to obtain a medical enhancement histogram; The medical enhancement histogram is filtered to determine a medical grayscale image.

4. A DR imaging device, characterized in that: A medical imaging control unit is included, and the medical imaging control unit includes: An acquisition module is used to acquire and filter the patient's scanned image to obtain a medical grayscale image; A processing module, configured to perform depth reconstruction on the medical grayscale image to obtain a display depth image, obtain a pathology decision value, perform decision segmentation on the display depth image according to the pathology decision value, and obtain a pathology feature image set; The processing module is further used to perform imaging dimension conversion on the pathological feature image set to obtain a medical model recognition matrix, and determine a medical recognition factor according to the medical model recognition matrix; The processing module is further used to determine the feature accumulation operators corresponding to the pathological feature region images in the pathological feature image set, obtain a feature accumulation operator sequence, and perform accumulation operator truncation estimation according to the feature accumulation operator sequence to obtain a medical truncation value; The processing module is further used to perform perception screening on the pathological feature image set according to the medical cutoff value and the medical perception factor to determine the medical abnormal site information set; An imaging module, used for generating a medical imaging marker map according to the medical abnormality site information set; Obtaining pathological decision values ​​specifically includes: determining a medical pixel attribute value according to the displayed depth map; Obtain all medical high-frequency pixel components in the displayed depth map; Obtain all medical low-frequency pixel components in the displayed depth map; Determine a pathology decision value according to a medical high-frequency pixel component, a medical low-frequency pixel component and a medical pixel attribute value; The displayed depth map is determined and segmented according to the pathological decision value to obtain a pathological feature image set, which specifically includes: The displayed depth map is segmented and decided according to the pathological decision value to obtain a plurality of medical edge images; Binarize all medical edge images to obtain a set of pathological feature images; The imaging dimension conversion of the pathological feature image set is performed to obtain a medical model recognition perception matrix, which specifically includes: Acquire all pathological characteristic region images in a pathological characteristic image set; Determine the medical feature matrix from all pathological feature area maps; Perform dimensionality reduction processing on the medical feature matrix to determine the medical model recognition matrix; The principal component analysis method is used to reduce the dimension of the medical feature matrix and determine the medical model recognition matrix; Among them, the pathological decision value can be determined according to the following formula: in, Represents the weighted average of medical high-frequency pixel components, Represents the weighted average of medical low-frequency pixel components, Represents the medical pixel attribute value, Represents the base of natural logarithms of Power; The medical model recognition matrix can be determined according to the following formula: in, represents the transposed matrix of the medical sample matrix, represents the medical feature matrix, represents the medical model recognition matrix; Among them, medical awareness factor Determined by the following formula: in, represents the maximum eigenvalue of the medical covariance matrix, represents the co-square calibration component, Represents medical pixel attribute value; Among them, the first Medical feature accumulation operator corresponding to the pathological feature region map It can be determined according to the following formula: in, represents the medical image recognition coefficient, represents the co-square calibration component, Represents the medical pixel attribute value, represents the medical image heterogeneity value, For the The pixel mean of the pathological feature area map; Among them, the medical cutoff value It can be determined by the following formula: in, Represents the total number of characteristic pixels in all pathological characteristic area maps, represents the medical pixel attribute value, A is the sequence length of the feature accumulation operator sequence, is the first in the medical feature accumulation operator sequence The medical feature accumulation operator corresponding to the pathological feature region map, For the The number of pixels in the pathological feature area map.

5. A computer device, characterized in that: The computer device includes a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the DR imaging method according to any one of claims 1 to 3.

6. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the DR imaging method according to any one of claims 1 to 3 is implemented.

Images