A medical image automatic analysis system and method

Abstract

The present invention relates to the technical field of medical image processing, and specifically provides a medical image automatic analysis system and method. The system includes a time series dynamic monitoring module, a lesion area weighted fusion module, a boundary energy optimization module, a multi-channel filtering processing module, and a time series feature construction module. In the present invention, through time series dynamic monitoring, significant change areas are screened to accurately locate the evolution trend of lesions, realizing the time series tracking analysis of medical images. With the help of weighted fusion processing, lesion data at different times are integrated to improve the accuracy of time series analysis and reduce data fluctuation errors. By combining gradient analysis and gray-scale contrast, the lesion boundary is optimized to exclude noise interference and make the lesion contour clearer. The multi-channel filtering screens different frequency features of the image, highlights the differential information of the lesion, reduces irrelevant interference, enhances feature expression, and constructs a complete time series analysis framework for lesions, providing a comprehensive image analysis basis for precision medicine.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and particularly relates to a medical image automatic analysis system and method. Background Art

[0002] The technical field of medical image processing involves the use of computer technology to automatically process, analyze, and interpret medical images. The core content of this field involves the application of technical means such as the recognition, enhancement, segmentation, registration, classification, and diagnosis of medical images. Medical image processing can assist doctors in disease detection, diagnostic support, treatment planning, etc. With the rapid development of computer vision and artificial intelligence technologies, medical image processing has been widely applied in the field of image diagnosis, especially achieving remarkable progress in automated and intelligent analysis. This field includes image processing algorithms, machine learning models, deep learning technologies, etc., aiming to improve the accuracy, efficiency, and application scope of medical images.

[0003] Among them, a medical image automatic analysis system refers to a system that uses computer-aided analysis to automatically identify and process medical images. This system mainly focuses on the analysis and interpretation of data in medical images, and specifically realizes functions such as preprocessing, feature extraction, and classification of medical images through image processing technology. The system includes processes such as image denoising, contrast enhancement, and segmentation of the target area, and then marks and analyzes the lesion area in the image. This patent theme completes the recognition and analysis of images in an automated manner, thereby supporting the auxiliary diagnosis function of medical images, reducing manual intervention, and improving the analysis accuracy.

[0004] Although the prior art can achieve automatic recognition and analysis, there are limitations in the dynamic monitoring of the lesion area. It mainly relies on the analysis of images at a single time point, lacks continuous tracking of the evolution trend of the lesion, and is difficult to accurately evaluate the development trend of the lesion. In the process of lesion boundary recognition, it is easily affected by image noise and uneven gray levels, resulting in blurred boundaries, affecting the precise positioning of the lesion area, and reducing the reliability of automatic analysis. In the existing image processing methods during the feature extraction process, they rely on filtering methods with fixed parameters and are difficult to take into account the specificity of different image data, resulting in some key features not being effectively highlighted and affecting the accuracy of image analysis. The recognition of the lesion area mainly relies on static images and fails to fully utilize time-series image data for multi-dimensional analysis, making it difficult to comprehensively evaluate the dynamic evolution characteristics of the lesion during the diagnosis process and affecting the accurate judgment of the disease process. In the process of automatic interpretation of medical images, the prior art has not established an effective time-series feature analysis framework, unable to fully reflect the development trend of the lesion, and limiting the clinical guiding value of image analysis. Summary of the Invention

[0005] The object of the present invention is to solve the deficiencies existing in the prior art, and to propose a medical image automatic analysis system and method.

[0006] To achieve the above object, the present invention adopts the following technical solutions: A medical image automatic analysis system includes:

[0007] The time series dynamic monitoring module is based on medical image data at different time points, extracts the volume, shape, and gray value features of the lesion area, calculates the change rate of the lesion area between adjacent time points, screens the image sequences with prominent change rates, judges the time series change features in the medical image data, and obtains the dynamic change trend of the lesion;

[0008] The weighted fusion module of the lesion area is based on the dynamic change trend of the lesion, extracts the change rate of the lesion at each time point, calculates the weight value of the change rate of the lesion in the image time series, screens the time series image data, and obtains the weighted fusion lesion area data;

[0009] The boundary energy optimization module is based on the weighted fusion lesion area data, identifies the gray contrast value between the lesion area and the surrounding tissues, analyzes the energy distribution of the lesion area boundary, screens the boundary features of the area affected by noise in the medical image data, and obtains the optimized lesion area boundary;

[0010] The multi-channel filtering processing module is based on the optimized lesion area boundary, extracts the frequency distribution information in the medical image data, screens the lesion area features in different frequency ranges, performs filtering processing on the medical image, and obtains the multi-channel filtering lesion features.

[0011] As a further solution of the present invention, the dynamic change trend of the lesion includes the volume change trend, the shape change trend, and the gray value change trend. The weighted fusion lesion area data includes weighted volume data, weighted shape data, and weighted gray data. The optimized lesion area boundary includes a denoised boundary, a contrast-enhanced boundary, and a smoothed boundary. The multi-channel filtering lesion features include high-frequency features, low-frequency features, and medium-frequency features.

[0012] As a further solution of the present invention, the time series dynamic monitoring module includes:

[0013] The lesion area feature recognition sub-module is based on medical image data at different time points, extracts the segmentation information of the lesion area, identifies the volume, shape, and gray value features of the lesion area, and statistically analyzes the pixel distribution parameters of the lesion area at each time point to obtain the multi-dimensional feature data of the lesion area;

[0014] The change rate calculation sub-module calls the multi-dimensional feature data of the lesion area, extracts the change values of the lesion area features at adjacent time points, and uses the formula:

[0015] ;

[0016] Calculate the change rate of the lesion area;

[0017] Wherein, represents the change rate of the lesion area, represents the time point lesion area volume at, represents the time point lesion area volume at, represents the time point lesion area shape feature parameter at, represents the time point lesion area shape feature parameter at, represents the time point lesion area gray value feature at, represents the time point lesion area gray value feature at;

[0018] The dynamic change pattern screening sub-module calls the change rate of the lesion area, screens the time points with prominent change rates, extracts the time series of the lesion areas with prominent changes, compares the trends of the lesion areas at different time points, and obtains the dynamic change trend of the lesion.

[0019] As a further solution of the present invention, the lesion area weighted fusion module includes:

[0020] The lesion rate calculation sub-module extracts the lesion area volume and edge contour change data in the image time series based on the dynamic change trend of the lesion, identifies the lesion change rate at each time point, adjusts the rate in combination with the lesion evolution trend, and uses the formula:

[0021] ;

[0022] Calculate the lesion change rate;

[0023] Wherein, represents the lesion change rate at the th time point, represents the lesion area volume at the th time point, represents the lesion area volume at the th time point, represents the real-time time corresponding to the th time point, represents the real-time time corresponding to the th time point, represents the th time point, the th edge contour point of the lesion area, the the lesion area of the n-th edge contour point at a time point, representing the total number of lesion edge contour points, representing the lesion shape adjustment coefficient;

[0024] The image data screening sub-module calls the lesion change rate to screen the image time series data, eliminates the image data with a change rate lower than the threshold, and obtains the screened image time series data;

[0025] The lesion area fusion sub-module, based on the screened image time series data, identifies the weights of the lesion change rate, and performs weighted fusion on the lesion area according to the weights to obtain the weighted fusion lesion area data.

[0026] As a further solution of the present invention, the boundary energy optimization module includes:

[0027] The boundary gradient extraction sub-module calculates the gray gradient in the image data based on the weighted fusion lesion area data, identifies the boundary gradient change rate, analyzes the local gradient change of the pixel gray value within the region, screens the boundary gradient mutation points, and obtains the boundary gradient distribution data;

[0028] The lesion area contrast analysis sub-module calls the boundary gradient distribution data, compares the gray levels of the lesion area and the surrounding tissues, and uses the formula:

[0029] ;

[0030] Calculate the gray contrast index of the lesion area boundary, screen the boundary area, and obtain the lesion area contrast feature;

[0031] where, represents the gray contrast index of the lesion area boundary, represents the gray value of the n-th pixel within the boundary gradient region, represents the average gray value of the pixels within the lesion area , represents the average gray value of the pixels within the surrounding tissue area , represents the number of pixel points within the boundary gradient region;

[0032] The noise influence screening sub-module calls the lesion area contrast feature, analyzes the noise influence area, identifies the energy distribution of the lesion area boundary, screens the abnormal areas in the energy distribution, eliminates the noise influence boundary, and obtains the optimized lesion area boundary.

[0033] As a further solution of the present invention, the multi-channel filtering processing module includes:

[0034] The spatial frequency analysis sub-module extracts pixel gray values based on the optimized lesion area boundary, analyzes the gray change trend at different positions, extracts spatial frequency information, determines the frequency distribution ratio, and uses the formula:

[0035] ;

[0036] Calculate the frequency distribution weight value of the pixel point to obtain the frequency distribution information;

[0037] Among them, represents the frequency distribution weight value of the pixel point, represents the pixel point gray value, represents the global gray mean value of the image, represents the total number of pixel points in the image, represents the local weight coefficient of the pixel point and represents the index of the pixel point;

[0038] The lesion area recognition sub-module calls the frequency distribution information, compares the frequency characteristics of the image area, filters out the frequency abnormal area, identifies the frequency deviation, and obtains the lesion area frequency characteristics;

[0039] The filtering feature recognition sub-module calls the lesion area frequency characteristics, adjusts the weight of the image frequency component, performs filtering processing, strengthens and weakens the frequency information of the key area, and obtains the multi-channel filtering lesion characteristics.

[0040] As a further solution of the present invention, the system further includes a timing feature construction module:

[0041] The timing feature construction module identifies the lesion area information at different time points in the medical image based on the multi-channel filtering lesion characteristics, extracts the key lesion characteristics in the image sequence, and obtains the lesion timing image analysis data;

[0042] The lesion timing image analysis data includes key lesion information, change pattern, and timing feature distribution.

[0043] As a further solution of the present invention, the timing feature construction module includes:

[0044] The filtered lesion signal calculation sub-module analyzes the signal change of the lesion area based on the multi-channel filtering lesion characteristics, calculates the image gradient value, filters out the lesion response channel, and obtains the lesion signal gradient value;

[0045] The lesion time point determination sub-module identifies the gradient change rate of the time series image based on the lesion signal gradient value, analyzes the time increment of the lesion area, and obtains the lesion time series change amount;

[0046] The lesion feature extraction sub-module identifies the morphological features of the lesion area based on the change amount of the lesion time series, analyzes the boundary changes at time points, and uses the formula:

[0047] ;

[0048] Calculate the characteristic change value of the lesion area to obtain the analysis data of the lesion time series image;

[0049] Among them, represents the characteristic change value of the lesion area, represents the gradient change amount of adjacent time points of the lesion area, represents the pixel offset of the lesion area at adjacent time points, represents the change rate of the lesion feature vector at adjacent time points, represents the lesion boundary change amount between adjacent time points.

[0050] The medical image automatic analysis method is executed based on the above medical image automatic analysis system, and includes the following steps:

[0051] S1: Based on the medical image data of different time points, extract the lesion area information in the image, calculate the volume, morphology, and gray value of the lesion area, arrange the data in chronological order, compare the lesion area characteristics of adjacent time points, identify the lesion change rate, and screen the image sequence with the change rate;

[0052] S2: Based on the image sequence with significant change rate, extract the lesion change rate of each time point, calculate the rate weight of the time point, select the optimal time point image data, and combine the lesion areas of multiple time points to obtain the weighted fusion lesion area data;

[0053] S3: Based on the weighted fusion lesion area data, extract the boundary gradient information in the image, calculate the gray contrast value between the lesion area and the surrounding tissues, screen the boundary points with gray contrast, analyze the distribution of the boundary points in the image, analyze the energy distribution of the regional boundary, eliminate the low-energy boundary points affected by noise, and obtain the optimized lesion area boundary;

[0054] S4: Based on the optimized lesion area boundary, extract the frequency distribution information in the image, screen the lesion area characteristics in different frequency ranges, filter the image data, and obtain the multi-channel filtered lesion characteristics;

[0055] S5: Based on the multi-channel filtered lesion characteristics, compare the lesion areas of different time series images, identify the key morphological features of the lesion area, extract the key point information of the lesion area at different time points, and obtain the analysis data of the lesion time series image.

[0056] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0057] In the present invention, through dynamic monitoring of time series, characteristic information such as the volume, shape, and gray scale of the lesion area is extracted, the change rate between adjacent images is calculated in the order of the time axis, the areas with significant change amplitude are screened, the dynamic evolution trend of the lesion is accurately positioned, and the time-series tracking analysis of medical images is realized. With the help of weighted fusion processing, based on the weight calculation of the lesion change rate, the lesion data at different times are comprehensively processed, the accuracy of image time-series analysis is improved, the error caused by data fluctuation is reduced, and for the boundary information of the lesion area, gradient analysis and gray-scale comparison means are used to optimize the area contour, eliminate noise interference, improve the edge recognition ability of the lesion area, and make the lesion morphology clearer. Through multi-channel filtering technology, different frequency characteristics in the image are screened, the differential information of the lesion area is highlighted, irrelevant interference in the image is effectively reduced, the expression ability of target lesion characteristics is enhanced, key characteristics are extracted from the image time-series data, a complete lesion time-series image analysis framework is constructed, the automatic interpretation ability of medical images is improved, the evolution trend of the lesion area is made more intuitive, and a more comprehensive image analysis basis is provided for precision medicine. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 is the system flow chart of the present invention;

[0059] Figure 2 is the flow chart of the time-series dynamic monitoring module in the present invention;

[0060] Figure 3 is the flow chart of the weighted fusion module of the lesion area in the present invention;

[0061] Figure 4 is the flow chart of the boundary energy optimization module in the present invention;

[0062] Figure 5 is the flow chart of the multi-channel filtering processing module in the present invention;

[0063] Figure 6 is the flow chart of the time-series feature construction module in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0065] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more, unless otherwise specifically defined.

[0066] Please refer to Figure 1 , the present invention provides a technical solution: A medical image automatic analysis system includes:

[0067] The time series dynamic monitoring module extracts the volume, shape, and gray value features of the lesion area based on the medical image data at different time points, arranges the data according to the time sequence, calculates the change rate of the lesion area between adjacent time points, screens the image sequences with prominent change rates, compares the dynamic change patterns of the lesion area, judges the time series change characteristics in the medical image data, and obtains the dynamic change trend of the lesion;

[0068] The weighted fusion module of the lesion area extracts the change rate of the lesion at each time point based on the dynamic change trend of the lesion, calculates the weight value of the change rate of the lesion in the image time series, screens the time series image data, and obtains the weighted fusion lesion area data;

[0069] The boundary energy optimization module extracts the boundary gradient information in the medical image data based on the weighted fusion lesion area data, identifies the gray contrast value between the lesion area and the surrounding tissues, analyzes the energy distribution of the lesion area boundary, screens the boundary characteristics of the area affected by noise in the medical image data, and obtains the optimized lesion area boundary;

[0070] The multi-channel filtering processing module extracts the frequency distribution information in the medical image data based on the optimized lesion area boundary, screens the lesion area characteristics in different frequency ranges, performs filtering processing on the medical image, and obtains the multi-channel filtered lesion characteristics;

[0071] The time series feature construction module identifies the lesion area information at different time points in the medical image based on the multi-channel filtered lesion characteristics, extracts the key lesion characteristics in the image sequence, and obtains the lesion time series image analysis data.

[0072] The dynamic change trends of lesions include volume change trends, shape change trends, and gray value change trends. The weighted fusion lesion area data includes weighted volume data, weighted shape data, and weighted gray data. The optimized lesion area boundary includes a denoised boundary, a contrast-enhanced boundary, and a smoothed boundary. The multi-channel filtered lesion features include high-frequency features, low-frequency features, and medium-frequency features. The lesion time-series image analysis data includes key lesion information, change patterns, and time-series feature distributions.

[0073] Please refer to Figure 2 , and the time-series dynamic monitoring module includes:

[0074] The lesion area feature recognition sub-module extracts the segmentation information of the lesion area based on medical image data at different time points, identifies the volume, shape, and gray value features of the lesion area, and statistically calculates the pixel distribution parameters of the lesion area at each time point to obtain the multi-dimensional feature data of the lesion area.

[0075] Collect medical image data at different time points through a high-precision scanner. For example, in a head MRI examination, continuously scan the lesion area of a brain tumor at different time points to collect a series of images. Through image processing software and edge detection algorithms, segment the lesion area from the images at each time point, identify its boundary. This process involves using pixel-level comparison and gray value threshold segmentation techniques to ensure accurate segmentation. Quantitatively measure the volume and shape of the segmented lesion area. The volume is obtained by converting the number of pixels to the known scanning volume ratio, and the shape features are described by calculating geometric parameters such as the curvature radius of the boundary line. The gray value feature is calculated by statistically averaging the gray values of all pixels in the lesion area and the standard deviation of their distribution. Compare the data with the patient's historical image data to analyze the change trend of the lesion area over time, and finally obtain the multi-dimensional feature data of the lesion area.

[0076] The change rate calculation sub-module calls the multi-dimensional feature data of the lesion area, extracts the change values of the lesion area features at adjacent time points, and uses the formula:

[0077] ;

[0078] Calculate the change rate of the lesion area;

[0079] Among them, represents the change rate of the lesion area, represents the volume of the lesion area at time point , represents the volume of the lesion area at time point , represents the shape feature parameter of the lesion area at time point , represents the shape feature parameter of the lesion area at time point ; The gray value characteristics of the lesion area representing the time point ; The gray value characteristics of the lesion area representing the time point ;

[0080] Calculate and compare the volume, shape, and gray value at consecutive time points to obtain the change rate of the lesion area. For example, assume that in the MRI scan of a certain patient at the first time point ( ), the volume of the lesion area is , the shape feature quantity (such as the comprehensive boundary curvature value) is , and the average gray value is . In the MRI scan at the second time point ( ), the volume of the lesion area grows to , the shape feature changes to , and the average gray value changes to ;

[0081] Calculate the change rate of the lesion area and substitute the specific values into the formula for calculation:

[0082] ;

[0083] The calculated change rate of the lesion area is 1.50, indicating that within this time interval, the volume of the lesion area has increased significantly. Considering the changes in shape features and gray values, the change rate exceeds the threshold of 1.0. Therefore, this time point can be marked as a time point with obvious lesion changes and further analysis of its clinical significance is required;

[0084] The innovation of the formula for calculating the change rate of the lesion area lies in that by comprehensively considering the volume change, shape feature change, and gray value change, it avoids the limitations brought by single - volume calculation, can more accurately reflect the dynamic change trend of the lesion area, improves the sensitivity to the development of the lesion, and the calculation result can be directly used to judge whether the lesion area has entered the abnormal growth stage, helping doctors decide whether further intervention or treatment plan adjustment is needed.

[0085] The dynamic change pattern screening sub - module calls the change rate of the lesion area, screens out the time points with prominent change rates, extracts the time series of the lesion area with prominent changes, compares the trends of the lesion area at different time points, and obtains the dynamic change trend of the lesion;

[0086] Identify time points with a significantly high rate of change through threshold screening technology. For example, in the case of periodic CT examinations of cancer patients, by comparing the rates of change in consecutive examinations, time points with abnormal rates are screened. The screening of time points is based on a preset clinical significance threshold. For instance, a rate exceeding 0.2 is considered clinically significant, and immediate treatment or further examinations are required. During the screening process, statistical analysis is performed on the data for each time point to compare the rate difference with the previous time point, in order to identify the dynamic change pattern of the lesion area. By comparing and analyzing the data at different time points, the response pattern of the lesion area to the progress of treatment and the real-time feedback of the treatment effect can be revealed, and ultimately the dynamic change trend of the lesion is obtained.

[0087] Please refer to Figure 3 , the weighted fusion module for the lesion area includes:

[0088] Based on the dynamic change trend of the lesion, the lesion rate calculation sub-module extracts the data on the volume and edge contour changes of the lesion area in the image time series, identifies the lesion change rate at each time point, and adjusts the rate in combination with the lesion evolution trend, using the formula:

[0089] ;

[0090] Calculate the lesion change rate;

[0091] Among them, represents the lesion change rate at the th time point, represents the volume of the lesion area at the th time point, represents the volume of the lesion area at the th time point, represents the real-time time corresponding to the th time point, represents the real-time time corresponding to the th time point, represents the lesion area of the th time point at the th edge contour point, the th time point at the th edge contour point of the lesion area, represents the total number of lesion edge contour points, represents the lesion shape adjustment coefficient;

[0092] The calculation logic of the "lesion change rate" is completed within the time series dynamic monitoring module. First, the system extracts features such as the volume, shape, and gray value of the lesion area based on medical image data at different times. By comparing the features of the lesion area between adjacent time points, the change rate of the lesion area is calculated. Specifically, the lesion change rate is evaluated by comprehensively considering the volume change, shape feature change, and gray value change of the lesion area. In the calculation formula, the lesion change rate is deduced from the differences in the lesion volume, shape parameters, and gray features at each time point, so as to obtain the change situation of the lesion in different time periods. This multi-dimensional calculation method can avoid the errors that may be brought by a single feature, and thus more accurately reflect the dynamic evolution trend of the lesion;

[0093] First, obtain the volume of the lesion area at each time point. The data is obtained from MRI or CT scans and quantified as volume values by relevant software. Suppose the MRI scan of a certain patient shows that the volume of the lesion area at the first time point is 200, and the volume of the lesion area at the second time point is 210, with a time interval of 24 hours (1 day). Then, calculate the lesion change rate by dividing the change difference in the volume of the lesion area between adjacent time points by the time difference, denotes the time interval between two adjacent time points, , ;

[0094] Substitute the values for calculation. First, calculate the first part, that is, the lesion volume change rate: ;

[0095] Secondly, consider the change in the edge contour of the lesion. Suppose there are 10 edge points, and the average movement of each point in one day is 0.2, 0.18, 0.22, 0.19, 0.21, 0.20, 0.23, 0.21, 0.19, 0.20. Then the second item is calculated as follows:

[0096] ;

[0097] Suppose the lesion shape adjustment coefficient is set to 0.5 (adjusted according to the lesion type and tissue morphology). Then the finally calculated lesion change rate is: ;

[0098] This value indicates that at the second time point, the volume growth rate of the lesion area plus the change in the edge contour makes the overall lesion change rate reach 10.102, which can better reflect the dynamic evolution trend of the lesion than simply calculating the volume change rate of 10. This calculation method ensures that the morphological changes of the lesion are incorporated into the analysis, making the data more comprehensive and accurate.

[0099] The image data screening sub-module calls the lesion change rate to screen the image time series data, eliminates the image data with a change rate lower than the threshold, and obtains the screened image time series data;

[0100] Set a threshold for the lesion change rate. For example, set the threshold to 0.5, which means that all time points with a lesion change rate lower than this value will be considered as having slow changes and will not be included in the key analysis sequence. The setting of this threshold is based on the analysis of historical case data. By statistically analyzing the distribution of the lesion change rate, the rate threshold that can reflect the significant progression of the lesion is determined. Then, the lesion change rate at each time point is compared, and the time points with a change rate higher than the threshold are screened out. The data at the screened time points will be used for further analysis. For example, in a monthly periodic liver scan, the three time points with the highest change rates are selected, and the corresponding image data will be used for the next weighted fusion analysis to ensure that only the most representative data is used for the final lesion analysis, thereby improving the accuracy and efficiency of the analysis and obtaining the screened image time series data.

[0101] The lesion area fusion sub-module, based on the screened image time series data, identifies the weights of the lesion change rate and performs weighted fusion on the lesion areas according to the weights to obtain the weighted fusion lesion area data;

[0102] Calculate the weight values of the lesion change rate at each time point in the image time series. The weight calculation is based on the proportion of the lesion change rate in the total rate change. Assume that the change rate at a certain time point accounts for 10% of the total rate, and its weight is also 10%. This weighting method ensures that the focus of the analysis is on the time points with the most significant changes. Perform weighted fusion on the lesion areas according to the weights. For example, if the weight of a certain time point is 30%, then the influence of the lesion image at this time point in the fused image is 30%. In this way, the weighted fusion lesion area data is obtained, which can more realistically reflect the overall development process of the lesion and provides an important visual basis for the diagnosis and treatment plan formulation of the lesion.

[0103] Please refer to Figure 4 , the boundary energy optimization module includes:

[0104] The boundary gradient extraction sub-module, based on the weighted fusion lesion area data, calculates the gray gradient in the image data, identifies the boundary gradient change rate, analyzes the local gradient change of the pixel gray values within the region, screens out the boundary gradient mutation points, and obtains the boundary gradient distribution data;

[0105] Select the lesion area from the medical image and calculate the change in the gray value of each pixel in this area relative to its surrounding pixels. Identify the approximate boundary of the lesion area in the image through image processing techniques, and then perform per-pixel analysis on the gray values of the boundary points. For example, in a 128*128 image, select a 10*10 area of the boundary region and calculate the gray difference between each pixel and its adjacent pixels. In this way, the gradient value of each point on the boundary can be obtained, and then the points with relatively large gradient mutations are selected as significant boundary feature points. The data will be used for subsequent lesion area identification and analysis. For instance, when processing a brain CT image, if it is found that the gradient value in a certain area suddenly jumps from 10 to 50, this indicates that there is a lesion or an important feature at the boundary. Obtain the boundary gradient distribution data, which can be directly used for the next step of lesion area comparison and analysis.

[0106] The lesion area comparison and analysis sub-module calls the boundary gradient distribution data, compares the gray levels of the lesion area and the surrounding tissues, and uses the formula:

[0107] ;

[0108] Calculate the gray contrast index of the lesion area boundary, screen the boundary area, and obtain the lesion area comparison features;

[0109] Among them, represents the gray contrast index of the lesion area boundary, represents the gray value of the th pixel in the boundary gradient area, represents the average gray value of the pixels within the lesion area , represents the average gray value of the pixels within the surrounding tissue area , represents the number of pixel points within the boundary gradient area;

[0110] The calculation logic of the gray contrast index of the lesion area boundary is based on the evaluation of the gray difference between the lesion area and the surrounding tissues. The specific process is to first calculate the average gray values of the lesion area and its surrounding normal tissue areas, and calculate the contrast through the difference in gray values. In the formula, by comparing the gray values of the lesion area and the surrounding tissue area, the gray contrast index is calculated to measure the degree of gray difference between the lesion area and the surrounding tissues. This calculation method involves comparing the gray value of each pixel within the lesion area with the average gray value of the surrounding tissue pixels, and the result is used to determine whether the boundary of the lesion area is obvious. In this calculation, the units of the gray values in the formula mentioned in the specification are not completely unified, resulting in a situation of inconsistent units, which may affect the accuracy of the calculation results;

[0111] First, determine the gray-scale contrast value between the lesion area and the surrounding tissues. This process requires calculating the average gray-scale values of the lesion area and its adjacent normal tissues, and measuring the contrast intensity through a normalization method. Assume a medical image of 512×512 pixels, the size of the pixel point set in the lesion area is m = 400, and the size of the pixel point set in the surrounding normal tissues is n = 1000. Obtain the gray-scale values of all pixel points in the lesion area and calculate their mean value: ;

[0112] Assume that the calculated average gray-scale value of the lesion area is , and perform the same calculation for the pixel points in the surrounding normal tissue area: ;

[0113] Assume that the calculated average gray-scale value of the surrounding tissues is , and calculate the gray-scale contrast index between the lesion area and the surrounding tissues ;

[0114] For actual calculation, we select some pixel values in the lesion area. For example, the gray-scale values of the pixel points in the lesion area are successively: 170, 190, 175, 185, 200. Then calculate the numerator part:

[0115] ;

[0116] Calculate the denominator part:

[0117] ;

[0118] Take the square root: ;

[0119] Finally, calculate the obtained gray-scale contrast index: ;

[0120] This value represents the degree of gray-scale difference between the lesion area and the surrounding tissues. The larger the value, the more obvious the difference between the lesion area and the surrounding tissues, and the easier it is to be recognized. If the contrast threshold is set to 0.3, at this time exceeds the threshold, then it is determined that this area has significant lesion characteristics. If is lower than the threshold, it belongs to the noise area or the area with unclear boundaries. This calculation is finally used to generate the lesion area contrast feature, which can be used for further lesion area segmentation and feature extraction to assist doctors in accurate diagnosis.

[0121] The noise influence screening sub-module calls the lesion area contrast feature, analyzes the noise influence area, identifies the energy distribution at the boundary of the lesion area, screens the abnormal areas in the energy distribution, eliminates the noise influence boundary, and obtains the optimized lesion area boundary;

[0122] Calculate and analyze the energy distribution at the boundary of the lesion area, identify irregular noise points introduced by device noise or image processing algorithms, and by setting a noise threshold, such as setting the noise contrast threshold to 0.5. If the energy value of a certain area is lower than this threshold, it is regarded as a noise-affected area and eliminated. The calculation of the energy distribution is obtained by weighted averaging the gray values of the identified boundary points to obtain a global energy distribution map. This chart reflects the visual prominence of each area. For brain scan images, this can help identify error areas caused by noise effects, thereby accurately depicting the true boundary of the lesion area and obtaining the optimized boundary of the lesion area. This is a key result data in the entire diagnostic process and is decisive for further pathological analysis and subsequent treatment plans.

[0123] Please refer to Figure 5 , the multi-channel filtering processing module includes:

[0124] Based on the optimized boundary of the lesion area, the spatial frequency analysis sub-module extracts the pixel gray values, analyzes the gray change trend at different positions, extracts the spatial frequency information, and judges the frequency distribution ratio. The formula is used:

[0125] ;

[0126] Calculate the frequency distribution weight value of the pixel point to obtain the frequency distribution information;

[0127] Among them, represents the frequency distribution weight value of the pixel point, represents the pixel point gray value, represents the global gray mean value of the image, represents the total number of pixel points in the image, represents the pixel point local weight coefficient, represents the index of the pixel point;

[0128] Extract the pixel gray values from the image data. The process involves digitizing the original image to extract gray information. In practical applications, such as in cardiac ultrasound image analysis, the gray value of each pixel point represents the intensity of the echo signal. Doctors can use the data to judge the density change of myocardial tissue. For a certain cardiac ultrasound image, set the total number of pixels to , the global gray mean value is , select the pixel points in a certain area for analysis. For example, set the gray values of five pixel points included in this area to be , , , , , and set the local weight coefficients of the pixel points to be 、 、 、 、 ;

[0129] Step 1: Calculate the gray deviation value of each pixel point;

[0130] ;

[0131] ;

[0132] ;

[0133] ;

[0134] ;

[0135] Step 2: Calculate the denominator part (the square root of the sum of the squares of the gray value deviations);

[0136] ;

[0137] ;

[0138] Step 3: Calculate the normalized frequency value and multiply it by the corresponding weight ;

[0139] ;

[0140] ;

[0141] ;

[0142] ;

[0143] ;

[0144] Finally, calculate the frequency distribution weight value ;

[0145] This value represents the frequency distribution weight of this region. Doctors can compare its numerical size with the region to further determine whether there is abnormal gray change in this region. If this value deviates from the normal myocardial tissue range (such as the normal tissue range is between and ), it indicates that there is a lesion in this region, and further analysis of the morphological characteristics of the myocardium is needed to more accurately locate the lesion region.

[0146] The lesion area recognition sub-module calls the frequency distribution information, compares the frequency characteristics of the image area, screens the areas with abnormal frequencies, identifies the frequency deviation, and obtains the frequency characteristics of the lesion area.

[0147] Make a detailed comparison of different regions in the image, mainly involving the use of advanced image processing techniques to analyze the frequency characteristics of each region. Taking echocardiogram as an example, the lesion areas with abnormal frequencies can be screened by comparing the frequency characteristics of different cardiac regions. For the calculation of the frequency deviation value, doctors can set specific thresholds to determine which regions have frequency deviations beyond the normal range. The thresholds are set based on a large amount of clinical data and expert experience. The specific setting process involves statistical analysis of the frequency distribution of past cases, and based on this, reasonable judgment criteria are formulated. In this way, the lesion areas in the echocardiogram can be effectively identified, and the specific location and scope of the lesion areas can be further determined, thus providing decision support for subsequent treatment and obtaining the frequency characteristics of the lesion areas.

[0148] The filtering feature recognition sub-module calls the frequency characteristics of the lesion area, adjusts the weights of the frequency components of the image, performs filtering processing, strengthens and weakens the frequency information of key areas, and obtains multi-channel filtered lesion characteristics.

[0149] Adjust and apply the filtering processing technology. The processing mainly involves selecting appropriate filters to strengthen or weaken the frequency information of specific regions. For the actual example of echocardiogram images, if the frequency characteristics of a certain region indicate potential lesions, the weights of the frequency components in this region can be adjusted to enhance the diagnostic effect of the image. The selection and parameter setting of the filters are based on specific pathological characteristics and expected image effects. The specific setting process includes analyzing the frequency characteristics of the lesion area, selecting the filter type that can maximize the highlighting of the characteristics, and then adjusting the intensity of the filter according to the severity of the lesion. Such processing not only makes the lesion area more obvious in the image but also facilitates doctors to make more accurate evaluations. Finally, the multi-channel filtered lesion characteristics are obtained, which will directly affect the accuracy of the diagnosis result and the formulation of the treatment plan, providing a more accurate and personalized diagnostic tool for doctors.

[0150] Please refer to Figure 6 , the time series feature construction module includes:

[0151] The filtered lesion signal calculation sub-module analyzes the signal changes in the lesion area based on the multi-channel filtered lesion characteristics, calculates the image gradient value, screens the lesion response channels, and obtains the lesion signal gradient value.

[0152] Call the multi-channel filtering results to calculate the lesion signal. First, obtain the image data, perform filtering processing on different channels, calculate the lesion signal changes in each channel. For each channel, analyze its image gradient value, which involves performing differential operations on the gray values of each pixel point to find the area with the largest gray change. Then, compare the data and select the channel with the strongest lesion response. For example, for a hypothetical lung CT scan, the lesion areas under different channels can be compared to find the morphological changes of lung nodules, which involves determining the boundary between the lesion area and the non-lesion area through the gray value difference. Finally, calculate the signal intensity of the channel and analyze the correlation between it and the gradient change. This process includes using statistical software to perform regression analysis on the signal intensity and the gradient change amount to determine whether there is a significant correlation, so as to obtain the lesion signal gradient value.

[0153] The lesion time point determination sub-module, based on the lesion signal gradient value, identifies the gradient change rate of the time series images, analyzes the time increment of the lesion area, and obtains the lesion time series change amount.

[0154] First, calculate the lesion gradient change rate of each frame image in the time series, which includes comparing the gray values of the lesion areas between consecutive frames and calculating their change rate. Then, analyze the time increment of the lesion gradient change between image frames, that is, calculate the time interval of the gray gradient change of the lesion area from one frame to the next frame, compare the change thresholds between different frames, and screen out the time points with significant changes according to the set gray change threshold. The time points are determined by calculation. For example, in a liver MRI scan, the development stage of liver cirrhosis lesions can be calibrated. Finally, identify the time points with significant lesion gradient changes and calculate the change amplitude between the time points, which can be achieved by comparing the gray statistical distributions of the lesion areas at each time point to obtain the lesion time series change amount.

[0155] The lesion feature extraction sub-module, based on the lesion time series change amount, identifies the morphological features of the lesion area, analyzes the boundary changes at the time points, and uses the formula:

[0156] ;

[0157] Calculate the lesion area feature change value to obtain the lesion time series image analysis data.

[0158] Among them, represents the lesion area feature change value, represents the gradient change amount of adjacent time points in the lesion area, represents the pixel offset of the lesion area between adjacent time points, represents the change rate of the lesion feature vector between adjacent time points, represents the lesion boundary change amount between adjacent time points;

[0159] Starting from the change amount of the lesion time series, morphological features of the lesion area changing over time are calculated, which involves morphological analysis of the lesion area at each time point. For example, edge detection algorithms are used to determine the boundary changes of the lesion area, and the pixel distribution offset is calculated. For a specific example, assume that in consecutive MRI scan frames, the pixel position of a tumor moves from (100, 200) in the first frame to (110, 210) in the second frame, indicating a lateral offset of 10 pixels and a longitudinal offset of 10 pixels. By establishing a lesion feature vector matrix, which is composed of lesion feature vectors at each time point, representing the changing trend of the lesion morphology;

[0160] Assume the specific values are: , , , , , , , ;

[0161] Substitute into the formula for calculation: ;

[0162] The morphological change intensity of the lesion area between consecutive time points can be calculated. This value (2.62) represents the degree of characteristic change of the lesion area from one time point to another, which can be used to evaluate the treatment effect or the progression of the lesion, and obtain the data of lesion time-series image analysis.

[0163] The medical image automatic analysis method is performed based on the above medical image automatic analysis system, including the following steps:

[0164] S1: Based on the medical image data at different time points, extract the lesion area information in the image, calculate the volume, morphology, and gray value of the lesion area, arrange the data in chronological order, compare the lesion area characteristics at adjacent time points, identify the lesion change rate, and screen the image sequence of the change rate;

[0165] S2: Based on the image sequence with a significant change rate, extract the lesion change rate at each time point, calculate the rate weight of the time point, select the optimal time point image data, and combine the lesion areas at multiple time points to obtain the weighted fusion lesion area data;

[0166] S3: Based on the weighted fusion lesion area data, extract the boundary gradient information in the image, calculate the gray contrast value between the lesion area and the surrounding tissues, screen the boundary points of the gray contrast, analyze the distribution of the boundary points in the image, analyze the energy distribution of the regional boundary, eliminate the low-energy boundary points affected by noise, and obtain the optimized lesion area boundary;

[0167] S4: Based on the optimized boundary of the lesion area, extract the frequency distribution information in the image, screen the lesion area features in the differential frequency range, filter the image data, and obtain the multi-channel filtered lesion features;

[0168] S5: Based on the multi-channel filtered lesion features, compare the lesion areas of the differential temporal images, identify the key morphological features of the lesion areas, extract the key point information of the lesion areas at the differential time points, and obtain the lesion temporal image analysis data.

[0169] The above are only the preferred embodiments of the present invention, and do not limit the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution content of the present invention still belong to the protection scope of the technical solution of the present invention.

Claims

1. An automatic medical image analysis system, characterized in that, The system includes: Based on the medical image data at different time points, the time series dynamic monitoring module extracts the volume, shape, and gray value features of the lesion area, calculates the change rate of the lesion area between adjacent time points, screens the image sequences with prominent change rates, judges the time series change characteristics in the medical image data, and obtains the dynamic change trend of the lesion; Based on the dynamic change trend of the lesion, the weighted fusion module for the lesion area extracts the change rate of the lesion at each time point, calculates the weight value of the change rate of the lesion in the image time series, screens the time series image data, and obtains the weighted fusion lesion area data; Based on the weighted fusion lesion area data, the boundary energy optimization module identifies the gray contrast value between the lesion area and the surrounding tissues, analyzes the energy distribution of the lesion area boundary, screens the boundary feature of the area affected by noise in the medical image data, and obtains the optimized lesion area boundary; The boundary energy optimization module includes: Based on the weighted fusion lesion area data, the boundary gradient extraction sub-module calculates the gray gradient in the image data, identifies the boundary gradient change rate, analyzes the local gradient change of the pixel gray value in the area, screens the boundary gradient mutation points, and obtains the boundary gradient distribution data; The lesion area contrast analysis sub-module calls the boundary gradient distribution data, compares the gray levels of the lesion area and the surrounding tissues, and uses the formula: ； Calculate the gray contrast index of the lesion area boundary, screen the boundary area, and obtain the lesion area contrast feature; Among them, The gray contrast index representing the boundary of the lesion area, Represents the gray value of the th pixel in the boundary gradient area, Represents the lesion area The average gray value of the pixels inside, Represents the surrounding tissue area The average gray value of the pixels inside, Represents the number of pixel points in the boundary gradient area; The noise influence screening sub-module calls the lesion area contrast feature, analyzes the noise influence area, identifies the energy distribution of the lesion area boundary, screens the abnormal area in the energy distribution, eliminates the boundary affected by noise, and obtains the optimized lesion area boundary; Based on the optimized lesion area boundary, the multi-channel filtering processing module extracts the frequency distribution information in the medical image data, screens the lesion area features in different frequency ranges, performs filtering processing on the medical image, and obtains the multi-channel filtering lesion features.

2. The medical image automatic analysis system according to claim 1, wherein The dynamic change trend of the lesion includes the volume change trend, shape change trend, and gray value change trend. The weighted fusion lesion area data includes weighted volume data, weighted shape data, and weighted gray data. The optimized lesion area boundary includes a denoised boundary, a contrast-enhanced boundary, and a smoothed boundary. The multi-channel filtering lesion features include high-frequency features, low-frequency features, and medium-frequency features.

3. The medical image automatic analysis system according to claim 1, characterized in that The time series dynamic monitoring module includes: Based on the medical image data at different time points, the lesion area feature recognition sub-module extracts the segmentation information of the lesion area, identifies the volume, shape, and gray value features of the lesion area, and statistically analyzes the pixel distribution parameters of the lesion area at each time point to obtain the multi-dimensional feature data of the lesion area; The change rate calculation sub-module calls the multi-dimensional feature data of the lesion area, extracts the change value of the lesion area feature at adjacent time points, and uses the formula: ； Calculate the change rate of the lesion area; Among them, represents the change rate of the lesion area, represents the time point of the volume of the lesion area, represents the time point of the volume of the lesion area, represents the time point of the shape feature parameter of the lesion area, represents the time point of the shape feature parameter of the lesion area, represents the time point of the gray value feature of the lesion area, represents the time point of the gray value feature of the lesion area; The dynamic change pattern screening sub-module calls the change rate of the lesion area, screens the time points with prominent change rates, extracts the time series of the lesion area with prominent changes, compares the trends of the lesion area at different time points, and obtains the dynamic change trend of the lesion.

4. The medical image automatic analysis system according to claim 3, wherein The diseased area weighted fusion module includes: Based on the dynamic change trend of the lesion, the lesion rate calculation sub-module extracts the volume of the lesion area and the change data of the edge contour in the image time series, identifies the lesion change rate at each time point, adjusts the rate in combination with the lesion evolution trend, and uses the formula: ； The lesion change rate is calculated; Among them, represents the lesion change rate at the th time point, represents the volume of the lesion area at the th time point, represents the volume of the lesion area at the th time point, represents the real-time time corresponding to the th time point, represents the real-time time corresponding to the th time point, represents the lesion area of the th edge contour point at the th time point, the th time point the th edge contour point of the lesion area, represents the total number of lesion edge contour points, represents the lesion morphology adjustment coefficient; The image data screening sub-module calls the lesion change rate, screens the image time series data, eliminates the image data with a change rate lower than the threshold, and obtains the screened image time series data; Based on the screened image time series data, the diseased area fusion sub-module identifies the weight of the lesion change rate, and performs weighted fusion on the diseased area according to the weight to obtain the weighted fusion diseased area data.

5. The medical image automatic analysis system according to claim 1, wherein The multi-channel filtering processing module includes: Based on the optimized boundary of the lesion area, the spatial frequency analysis sub-module extracts the pixel gray value, analyzes the gray change trend at different positions, extracts the spatial frequency information, and judges the frequency distribution ratio. The formula is: ； The frequency distribution weight value of the pixel point is calculated to obtain the frequency distribution information; Among them, represents the frequency distribution weight value of the pixel point, represents the pixel point gray value, represents the global gray mean of the image, represents the total number of pixel points in the image, represents the pixel point local weight coefficient, represents the index of the pixel point; The diseased area identification sub-module calls the frequency distribution information, compares the frequency characteristics of the image area, screens the areas with abnormal frequencies, identifies the frequency deviation, and obtains the frequency characteristics of the diseased area; The filtering feature identification sub-module calls the frequency characteristics of the diseased area, adjusts the weight of the image frequency component, performs filtering processing, strengthens and weakens the frequency information of the key area, and obtains the multi-channel filtering diseased characteristics.

6. The medical image automatic analysis system according to claim 1, wherein The system further includes a timing feature construction module: Based on the multi-channel filtering diseased characteristics, the timing feature construction module identifies the diseased area information at different time points in the medical image, extracts the key diseased characteristics in the image sequence, and obtains the diseased timing image analysis data; The diseased timing image analysis data includes key diseased information, change patterns, and timing feature distributions.

7. The medical image automatic analysis system according to claim 6, characterized in that The timing feature construction module includes: Based on the multi-channel filtering diseased characteristics, the filtered diseased signal calculation sub-module analyzes the signal change of the diseased area, calculates the image gradient value, screens the diseased response channels, and obtains the diseased signal gradient value; Based on the diseased signal gradient value, the diseased time point determination sub-module identifies the gradient change rate of the time series image, analyzes the time increment of the diseased area, and obtains the diseased time series change amount; Based on the diseased time series change amount, the diseased feature extraction sub-module identifies the morphological features of the diseased area, analyzes the boundary change at the time point, and uses the formula: ； The diseased area feature change value is calculated to obtain the diseased timing image analysis data; Among them, represents the characteristic change value of the lesion area, represents the gradient change amount of adjacent time points of the lesion area, represents the pixel offset of the lesion area at adjacent time points, represents the change rate of the lesion feature vector at adjacent time points, represents the change amount of the lesion boundary between adjacent time points.

8. An automatic medical image analysis method, characterized in that, Executed by the medical image automatic analysis system according to any one of claims 1-7, including the following steps: S1: Based on the medical image data at different time points, extract the diseased area information in the image, calculate the volume, morphology, and gray value of the diseased area, arrange the data in chronological order, compare the diseased area characteristics of adjacent time points, identify the diseased change rate, and screen the image sequence of the change rate; S2: Based on the image sequence with a significant change rate, extract the lesion change rate at each time point, calculate the rate weight of the time point, select the optimal time point image data, and combine the lesion regions at multiple time points to obtain the weighted fusion lesion region data; S3: Based on the weighted fusion lesion region data, extract the boundary gradient information in the image, calculate the gray contrast value between the lesion region and the surrounding tissues, screen the boundary points of the gray contrast, analyze the distribution of the boundary points in the image, analyze the energy distribution of the regional boundary, eliminate the low-energy boundary points affected by noise, and obtain the optimized lesion region boundary; S4: Based on the optimized lesion region boundary, extract the frequency distribution information in the image, screen the lesion region features in the differential frequency range, filter the image data, and obtain the multi-channel filtered lesion features; S5: Based on the multi-channel filtered lesion features, compare the lesion regions of the differential time-series images, identify the key morphological features of the lesion regions, extract the key point information of the lesion regions at different time points, and obtain the lesion time-series image analysis data.

Images