Autism detection method and system based on brain function imaging data analysis
By constructing a shaking influence factor model and using adaptive histogram equalization technology to correct brain structure images, the problem of image loss or redundancy caused by shaking in autistic patients was solved, thus improving the accuracy of autism detection.
Patent Information
- Application Number
- CN202511312244.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-15
- Publication Date
- 2025-12-19
AI Technical Summary
Autistic patients are prone to shaking during brain functional imaging data acquisition, which can lead to missing or excessive brain structures and reduce the accuracy of image information.
By constructing a swaying influencing factor model, the degree of swaying in patients is quantified in real time. Combined with adaptive histogram equalization technology and time series contour matching algorithm, brain structure images are corrected to ensure the integrity and accuracy of image information.
This effectively avoids the loss or redundancy of brain structure images caused by limb movement, improves the integrity and accuracy of image information, and provides reliable data support for autism detection.
Smart Images
Figure CN121169867A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of image analysis, in particular to an autism detection method and system based on brain function imaging data analysis. BACKGROUND
[0002] The clinical diagnosis of autism currently mainly relies on behavior observation and scale evaluation, which has strong subjectivity, significant heterogeneity influence, and difficulty in accurately capturing early core symptoms, etc. Therefore, based on brain function imaging technology, the functional signals of the brain of an autism patient in a resting state or when performing a specific task are collected by a special medical device, external interference is removed after a preprocessing step, features such as brain region activation patterns and functional connection strength are extracted, and compared with the brain function data model of normal people, the neural mechanism abnormalities related to autism are identified, and finally the detection result is output through algorithm analysis to assist clinical diagnosis.
[0003] However, autism patients are sensitive and easily distracted, which causes the patient to unconsciously shake or move when collecting their brain images, resulting in missing or excessive brain structures in the images, and thus reducing the accuracy of image information. SUMMARY
[0004] Technical problems to be solved
[0005] In view of the deficiencies in the prior art, the present application provides an autism detection method and system based on brain function imaging data analysis, which solves the problem of missing or excessive brain structures caused by the movement of autism patients, which in turn reduces the accuracy of image information.
[0006] Technical scheme
[0007] To achieve the above purpose, the present application is implemented by the following technical scheme: an autism detection method and system based on brain function imaging data analysis, comprising the following specific steps and modules: step one: start real-time collection of patient images and brain structure images, and comprehensively analyze the patient images to obtain a shaking influence factor; step two: judge whether the patient shakes according to the shaking influence factor, if it is judged that the patient shakes, a prompt is issued, and the process returns to step one, if it is judged that the patient does not shake, step three is continued; step three: comprehensively analyze the brain structure images to obtain a brain structure abnormality factor; step four: judge whether the brain structure is abnormal according to the brain structure abnormality factor, if it is judged that the brain structure in the image is abnormal, the brain structure in the image is corrected through a brain structure correction algorithm, and step five is executed, if it is judged that the brain structure in the image is normal, the detection is ended; step five: judge whether the brain structure in the image is successfully corrected, if it is judged that the correction is successful, the process returns to step three, if it is judged that the correction fails, the process returns to step two.
[0008] Further, the specific acquisition method of the shaking influence factor is as follows: the contour movement parameter is obtained by comprehensive calculation on the patient image, the brightness fluctuation parameter is obtained by comprehensive calculation on the brain structure image, the shaking influence factor is obtained by standardization processing and comprehensive calculation according to the contour movement parameter and the brightness fluctuation parameter; HY = CC x LC; wherein, HY represents the shaking influence factor, CC represents the contour movement parameter, and LC represents the brightness fluctuation parameter.
[0009] Further, the specific acquisition method of the contour movement parameter is as follows: the pixel point coordinates of each pixel point in the patient image are obtained by establishing a three-dimensional coordinate system according to the patient image, the patient contour is obtained by processing the patient image through an image contour segmentation algorithm and an image contour tracking algorithm, the total coordinates are obtained by sum calculation on the pixel point coordinates in the patient contour according to the number of pixel points in the patient contour, the contour center point is obtained by quotient calculation on the total coordinates according to the number of pixel points in the patient contour, the movement distance and the movement angle of the contour center point are calculated according to a time sequence, the movement distance and the movement angle are obtained, the movement distance and the movement angle are standardized processed to remove the dimension and are comprehensively calculated to obtain the contour movement parameter.
[0010] Further, the specific acquisition method of the movement distance is as follows: in the time sequence, the movement distance is obtained by distance calculation on the contour center point of the next second and the contour center point of the last second through the Euclidean distance formula.
[0011] Further, the specific acquisition method of the brightness fluctuation parameter is as follows: the pixel point coordinates of each pixel point in the brain structure image are obtained by establishing a two-dimensional coordinate system according to the brain structure image, the brightness value of each pixel point is obtained by gray scale conversion on the brain structure image, and the brightness value of each pixel point is recorded as brightness, the balanced brightness of each frame is obtained by sum and average calculation on the brightness value of each pixel point in each frame of the brain structure image according to the number of pixel points in each frame of the brain structure image, the brightness deviation value of each pixel point is obtained by difference calculation and absolute value taking on the brightness value of each pixel point in each frame of the brain structure image with the balanced brightness of each frame in turn, the brightness confusion value of each frame of the brain structure image is obtained by sum calculation on the brightness deviation value of each pixel point according to the number of pixel points in each frame of the brain structure image, and the brightness fluctuation parameter is obtained by sum calculation on the brightness confusion value of each frame of the brain structure image according to the time sequence.
[0012] Further, the specific acquisition method of the brain structure abnormality factor is as follows: the brain structure image is processed by an image contour segmentation algorithm and an image contour tracking algorithm to obtain a brain contour, the brain contour of the brain structure image is comprehensively calculated to obtain a brain contour abnormality parameter, and the brain contour abnormality parameter and a brightness fluctuation parameter are standardized to remove the dimension and are comprehensively calculated to obtain the brain structure abnormality factor; NY = LC x JC; wherein, NY represents the brain structure abnormality factor, LC represents the brightness fluctuation parameter, and JC represents the brain contour abnormality parameter.
[0013] Further, the specific acquisition method of the brain contour abnormality parameter is as follows: the same kind of brain contour in different time sequences is sequentially corresponded to obtain a brain contour data set, the brain contour data set includes brain contours in various time sequences, the area of each brain contour in each time sequence is calculated, that is, the pixel point number in each brain contour is summed to obtain the brain contour area in each time sequence, the brain contour area in each time sequence is used to calculate the difference between the brain contour area in the next second and the brain contour area in the last second and take the absolute value to obtain a brain contour area deviation value greater than zero, the brain contour area deviation value is summed to obtain a brain contour area deviation total value, and the brain contour area deviation total values of different kinds of brain contours are summed according to the number of brain contour types to obtain the brain contour abnormality parameter.
[0014] Further, the specific steps of sequentially corresponding the same kind of brain contour in different time sequences are as follows: a first brain contour in a first time brain structure image is T1N1, a certain brain contour in a second time brain structure image is T2N j , the pixel point coordinates in T1N1 and T2N j are counted, then the pixel point coordinates in T1N1 and T2N j are sequentially traversed and matched, if the same pixel point coordinates are matched, the number is counted, if different pixel point coordinates are matched, the matching is continued until the pixel point coordinates in T1N1 are matched with the pixel point coordinates in T2N j , different matching similarities are obtained, the pixel point coordinates in T2N j are arranged in the order from small to large according to the sorting algorithm, the maximum matching similarity is obtained, T2N j corresponding to the maximum matching similarity is selected, and the same kind of brain contour in the first time brain structure image is sequentially corresponded to the same kind of brain contour in the second time brain structure image, and the same kind of brain contour in different time sequences is sequentially corresponded in the same way.
[0015] Further, the specific step of correcting the brain structure in the image by the brain structure correction algorithm is: the brightness of the brain structure image is processed by adaptive histogram equalization, so that the brightness of the brain structure image is finally equalized; a single frame is randomly selected for the brain structure image with zero brain profile abnormality parameter between adjacent time sequences, to obtain the corrected brain structure image.
[0016] Further, the system comprises: an image data acquisition module, a historical data storage module, an image data analysis module, a data processing and correction module, and a correction success judgment module; the image data acquisition module is used for acquiring patient images and brain structure images in real time; the historical data storage module is used for storing historical normal patient images and brain structure images and calculating a judgment threshold; the image data analysis module is used for calling the image data acquisition module and the historical data storage module, and comprehensively analyzing to obtain a shaking influence factor and a brain structure abnormality factor, for judging whether the patient shakes and whether the brain structure is abnormal; the data processing and correction module is used for calling the image data analysis module, issuing a prompt to the patient according to the judgment result of the patient shaking, and correcting and adjusting the brain structure image according to the judgment result of the brain structure abnormality; and the correction success judgment module calls the data processing and correction module to judge whether the brain structure image is corrected successfully, and re-calls the image data analysis module according to the judgment result of the correction failure.
[0017] Advantages
[0018] Compared with the prior art, the embodiments of the present application have at least the following advantages or beneficial effects:
[0019] 1. By constructing a shaking influence factor model, the shaking degree of the patient is quantified in real time, and when the shaking is detected to exceed the threshold, a prompt is immediately issued to guide the patient to keep stable, effectively avoiding the brain structure image missing or redundancy problem caused by limb movement, ensuring the integrity and accuracy of the image information from the data acquisition source, and providing a reliable basis for subsequent analysis.
[0020] 2. The image brightness is equalized by the adaptive histogram equalization technology, and the brain profile recognition is enhanced; combined with the time sequence profile matching algorithm, the stable frame with no deviation in profile area between adjacent frames is selected for framing correction, so as to eliminate the brain structure misplacement or distortion caused by shaking, and significantly improve the reliability of brain structure abnormality judgment, and provide more accurate image data support for autism detection.
[0021] Of course, implementing any product of the present application does not necessarily need to achieve all the advantages described above at the same time. BRIEF DESCRIPTION OF DRAWINGS
[0022] Figure 1For the present application: a flow chart of an autism detection method based on brain function imaging data analysis.
[0023] Figure 2 For the present application: a system structure diagram of an autism detection system based on brain function imaging data analysis. DETAILED DESCRIPTION
[0024] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.
[0025] It should be noted that, in this document, relationship terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or device.
[0026] As shown in Figure 1 The embodiments of the present application provide an autism detection method based on brain function imaging data analysis: comprising the following specific steps:
[0027] Step one: start real-time acquisition of patient images through the camera, the patient images are three-dimensional images, improve the accuracy of the patient images, filter and denoise the patient images, remove the redundant values and abnormal values of the patient images, which helps to improve the data quality of the patient images, establish a three-dimensional coordinate system of x-axis, y-axis and z-axis according to the patient images, obtain the pixel point coordinates of each pixel point in the patient images, process the patient images through image contour segmentation algorithm and image contour tracking algorithm, obtain the patient contour, the image contour segmentation algorithm, such as sobel algorithm, is used to divide the contour of the patient images, the image contour tracking algorithm, such as Lucas-Kanade optical flow method, is used to track the motion of the divided contour, since the pixel point coordinates of the patient images are known, the pixel point coordinates in the patient contour are summed according to the number of pixel points in the patient contour, and the total coordinates are obtained, and then the total coordinates are divided according to the number of pixel points in the patient contour, and the contour center point is obtained. The contour center point is used as a feature point for subsequent calculation, and the contour movement parameters are obtained by comprehensively calculating the movement of each frame contour center point in the time sequence;
[0028] The brain structure image is acquired in real time by the MRI, and the brain structure image is filtered and denoised to remove redundant values and abnormal values in the brain structure image, which helps to improve the data quality of the brain structure image. After the hydrogen protons of human tissues are arranged in a certain direction by the main magnetic field of the MRI, the k-space matrix is filled row by row using the frequency encoding gradient and the phase encoding gradient. The k-space matrix is the spatial frequency domain of the original signal, and the center area corresponds to the low-frequency tissue profile of the brain structure image, and the edge area corresponds to the high-frequency structural details of the brain structure image. Finally, the k-space data is converted into a brain structure image in which the voxel position and signal intensity are mapped through inverse Fourier transform. Since the brain structure image is usually a two-dimensional image, a two-dimensional coordinate system of a-axis and b-axis is established according to the brain structure image to obtain the pixel point coordinates of each pixel point in the brain structure image. The brain structure image is converted into grayscale to obtain the brightness value of each pixel point, denoted as brightness. However, the shaking of the patient causes the frequency encoding gradient and the phase encoding gradient to be misaligned, resulting in chaotic brightness of the brain structure image. Therefore, the brightness of the brain structure image is comprehensively calculated to obtain a brightness fluctuation parameter. The brain structure image is processed by an image contour segmentation algorithm and an image contour tracking algorithm to obtain a brain contour.
[0029] The contour movement parameter and the brightness fluctuation parameter are standardized to remove dimensions and comprehensively calculated to obtain a shaking influence factor.
[0030] HY = CC x LC.
[0031] Wherein, HY represents the shaking influence factor, reflecting whether the patient with autism shakes during brain function detection, CC represents the contour movement parameter, reflecting whether the patient's contour moves, and LC represents the brightness fluctuation parameter, reflecting whether the brightness in the brain structure image is chaotic.
[0032] The specific acquisition method of the contour movement parameter is as follows:
[0033] The moving distance and the moving angle of the contour center point are calculated according to the time sequence to obtain the moving distance and the moving angle. The moving distance and the moving angle are standardized to remove dimensions and comprehensively calculated to obtain the contour movement parameter.
[0034]
[0035] Wherein, CC represents the contour movement parameter, t represents time, YD i represents the moving angle of the i-th second, and YL irepresents the moving distance of the i-th second, k represents a positive real number, because the greater the moving angle or the moving distance, the greater the profile moving parameter, so the positive real number is to avoid the situation that when the moving angle is zero, the moving distance is not zero, but the profile moving parameter is zero, that is, when the moving angle is zero, only the moving distance can also reflect the change of the profile moving parameter, and not only indicates that the profile moving parameter is zero, when the moving distance is zero, the moving angle is not zero, which is not practical, that is, the situation that the patient profile rotates based on the profile center point does not exist, so when the profile center point does not move, the moving distance is zero, and the moving angle is also zero, which reflects that the patient is in a non-shaking state, and when the profile center point moves, the moving distance is zero, which is meaningless.
[0036] The specific acquisition method of the moving distance is as follows:
[0037] In the time sequence, the distance between the profile center point of the next second and the profile center point of the last second is calculated by the Euclidean distance formula to obtain the moving distance, that is, let M1(x1, y1, z1) and M2(x2, y2, z2), wherein M1 is the profile center point of the last second, and M2 is the profile center point of the next second, wherein YL represents the moving distance, x2 represents the x-axis coordinate of the profile center point of the next second, x1 represents the x-axis coordinate of the profile center point of the last second, y2 represents the y-axis coordinate of the profile center point of the next second, y1 represents the y-axis coordinate of the profile center point of the last second, z2 represents the z-axis coordinate of the profile center point of the next second, and z1 represents the z-axis coordinate of the profile center point of the last second.
[0038] The specific acquisition method of the moving angle is as follows:
[0039] The vector calculation is performed on M1 and M2 to obtain a displacement vector, that is, wherein is the displacement vector, so the modulus of the displacement vector is the moving distance, let the angle between the profile center point of the next second and the profile center point of the last second about the x-axis be x 夹 , let the angle between the profile center point of the next second and the profile center point of the last second about the y-axis be y 夹 , and let the angle between the profile center point of the next second and the profile center point of the last second about the z-axis be z 夹 , because wherein cosx 夹 represents the cosine of x 夹 , and YL represents the moving distance, that is, the modulus of the displacement vector, because the moving distance is in the denominator, so further indicating that when the profile center point moves, the moving distance is zero, which is meaningless, so Similarly x 夹 , y 夹and z 夹 are summed up between 0 degree and 180 degree, x 夹 , y 夹 and z 夹 to obtain the moving angle.
[0040] The specific acquisition method of the brightness fluctuation parameter is as follows:
[0041] The brightness values of each pixel point in each frame of brain structure image are summed and then averaged to obtain the equal brightness of each frame, the brightness values of each pixel point in each frame of brain structure image are sequentially subtracted from the equal brightness of each frame and the absolute values are taken to obtain the brightness deviation value of each pixel point, and the brightness deviation values of each pixel point are summed to obtain the brightness confusion value of each frame of brain structure image, and the brightness confusion values of each frame of brain structure image are summed to obtain the brightness fluctuation parameter.
[0042] Step two: judging whether the patient shakes according to the shaking influence factor, setting a shaking threshold, using the value of the patient not shaking obtained by analyzing and calculating the historical patient images and brain structure images to judge whether the patient shakes, and storing it into the database, comparing the shaking influence factor with the shaking threshold in real time, if the shaking influence factor is greater than the shaking threshold, it is judged that the patient shakes, a prompt is given, and the process returns to step one, if the shaking influence factor is less than or equal to the shaking threshold, it is judged that the patient does not shake, and step three is continued.
[0043] Step three: comprehensively calculating the brain contour of the brain structure image to obtain a brain contour abnormality parameter, and standardizing the brain contour abnormality parameter and the brightness fluctuation parameter to remove the dimension and comprehensively calculating to obtain a brain structure abnormality factor.
[0044] NY = LC x JC.
[0045] Wherein, NY represents the brain structure abnormality factor, reflecting whether the brain structure in the brain structure image is abnormal, LC represents the brightness fluctuation parameter, when the brightness confusion in the brain structure image not only causes the image contour segmentation algorithm or the image contour tracking algorithm depending on the brightness value to fail to accurately segment or track the contour of the brain structure, but also causes the recognition degree of the brain contour segmented or tracked by the image contour segmentation algorithm or the image contour tracking algorithm to decrease, that is, the brain structure in the brain structure image is abnormal, JC represents the brain contour abnormality parameter, reflecting whether the brain contour is abnormal.
[0046] The specific acquisition method of the brain contour abnormality parameter is as follows:
[0047] The same kind of brain profile of different time sequences is sequentially corresponding to obtain a brain profile data set, the brain profile data set includes brain profiles of different time sequences, and the area of each brain profile of different time sequences is calculated sequentially, that is, the number of pixel points in each brain profile is summed to obtain the area of each brain profile of different time sequences, the area of the same kind of brain profile of the next second and the area of the same kind of brain profile of the last second are calculated by difference and the absolute value is taken according to the area of the brain profile of each different time sequence, to obtain a certain same kind of brain profile area deviation value greater than zero, the sum of the certain same kind of brain profile area deviation value is calculated to obtain the total value of the same kind of brain profile area deviation, and the total value of the different kinds of brain profile area deviation is calculated by sum according to the number of brain profile types to obtain the brain profile abnormality parameter.
[0048] The specific steps of sequentially corresponding the same kind of brain profile of different time sequences are as follows:
[0049] Let the first brain profile in the brain structure image at the first time be T1N1, and let a brain profile in the brain structure image at the second time be T2N j , the number of pixel points in T1N1 and T2N j is counted, then the pixel point coordinates in T1N1 and T2N j are sequentially traversed and matched, if the same pixel point coordinates are matched, the number is counted, if different pixel point coordinates are matched, the matching is continued until the pixel point coordinates in T1N1 are matched with the pixel point coordinates in T2N j , different matching similarities are obtained, that is, the greater the matching similarity, the more the number of pixel point coordinates in T1N1 and T2N j that are the same, the pixel point coordinates in T2N j are arranged in order from small to large according to the matching similarity by using a sorting algorithm to obtain the maximum matching similarity, T2N j corresponding to the matching similarity is selected, that is, T2N j is the first brain profile of T1N1 in the brain structure image at the second time, and the same kind of brain profile in the brain structure image at the first time is sequentially corresponding to the same kind of brain profile in the brain structure image at the second time, and the same kind of brain profile of different time sequences is sequentially corresponding.
[0050] Step four: According to the brain structure abnormality factor, whether the brain structure is abnormal is judged, the brain structure abnormality threshold is set, the normal value of the brain structure in the image obtained by analyzing and calculating the brain structure image collected in history is used to judge whether the brain structure in the image is normal, and is stored in the database. According to the real-time comparison between the brain structure abnormality factor and the brain structure abnormality threshold, if the brain structure abnormality factor is greater than the brain structure abnormality threshold, it is judged that the brain structure in the image is abnormal, the brain structure correction algorithm is used to correct the brain structure in the image, and step five is executed. If the brain structure abnormality factor is less than or equal to the brain structure abnormality threshold, it is judged that the brain structure in the image is normal, and the detection is ended.
[0051] The specific steps of correcting the brain structure in the image by the brain structure correction algorithm are:
[0052] The brightness of the brain structure image is processed by adaptive histogram equalization, that is, the brain structure image is divided into several regions, the brightness value of the region with high brightness is reduced, and the brightness value of the region with low brightness is increased. Finally, the brightness of the brain structure image is equalized, which is not only beneficial to the accurate segmentation or tracking of the outline of the brain structure by the image outline segmentation algorithm or the image outline tracking algorithm, but also restores the recognition degree of the brain outline and reduces the probability of brain structure abnormalities in the brain structure image;
[0053] When the brain structure image is normal, the brain outline in the brain structure image is also normal, that is, the brain structure in the brain structure image is not missing or excessive, so that the brain outline is not too small or too large, and the brain outline area under each different time sequence is equal, that is, the brain outline abnormality parameter is zero. Therefore, a single frame is randomly selected for the brain structure image with a brain outline abnormality parameter of zero between adjacent time sequences, to obtain a corrected brain structure image, so as to avoid the missing or excessive brain structure in the image, and further improve the accuracy of the image information.
[0054] Step five: The correction time period is set according to historical experience and stored in the database, which is used to judge whether the brain structure in the brain structure image is corrected successfully. If the brain outline abnormality parameters between adjacent time sequences in the correction time period are all zero, it is judged that the correction is successful, and step three is returned. If the brain outline abnormality parameters between adjacent time sequences in the correction time period are not all zero, it is judged that the correction fails, the single frame is cancelled, and step two is returned.
[0055] As shown in Figure 2 A system for detecting autism based on brain function imaging data analysis, comprising:
[0056] An image data acquisition module: used for real-time acquisition of patient images and brain structure images;
[0057] Historical data storage module: used for storing historical normal patient images and brain structure images and calculating a judgment threshold;
[0058] Image data analysis module: used for calling the image data acquisition module and the historical data storage module, and comprehensively analyzing to obtain a shaking influence factor and a brain structure abnormality factor, for judging whether the patient shakes and whether the brain structure is abnormal;
[0059] Data processing and correction module: used for calling the image data analysis module, issuing a prompt to the patient according to the judgment result of the patient shaking, and correcting and adjusting the brain structure image according to the judgment result of the brain structure abnormality;
[0060] Correction success judgment module: calling the data processing and correction module, judging whether the brain structure image is corrected successfully, and re-calling the image data analysis module according to the judgment result of the correction failure.
[0061] The preferred embodiments disclosed above are only used to help explain the present application. The preferred embodiments do not describe all the details and do not limit the present application to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of the present application. The present application is selected and described in detail to better explain the principles and practical applications of the present application, so that those skilled in the art can well understand and utilize the present application. The present application is limited by the claims and their entire scope and equivalents.
Claims
1. A method for autism detection based on brain function imaging data analysis, characterized in that: The method comprises the following specific steps: Step 1: start real-time acquisition of patient images and brain structure images, and comprehensively analyze the patient images to obtain a shaking influence factor; Step 2: judge whether the patient shakes according to the shaking influence factor, if it is judged that the patient shakes, a prompt is given, and the method returns to step 1, if it is judged that the patient does not shake, the method continues to step 3; Step 3: comprehensively analyze the brain structure images to obtain a brain structure abnormality factor; Step 4: judge whether the brain structure is abnormal according to the brain structure abnormality factor, if it is judged that the brain structure in the image is abnormal, correct the brain structure in the image through a brain structure correction algorithm, and execute step 5, if it is judged that the brain structure in the image is normal, end the detection; Step 5: judge whether the correction of the brain structure in the image is successful, if it is judged that the correction is successful, return to step 3, if it is judged that the correction fails, return to step 2. 2.The autism detection method based on brain function imaging data analysis of claim 1, characterized in that: The specific acquisition method of the shaking influence factor is as follows: Comprehensively calculate the patient images to obtain a contour movement parameter, comprehensively calculate the brain structure images to obtain a brightness fluctuation parameter, and comprehensively calculate the contour movement parameter and the brightness fluctuation parameter after standardization processing, to obtain the shaking influence factor; HY=CC×LC; Wherein, HY represents the shaking influence factor, CC represents the contour movement parameter, and LC represents the brightness fluctuation parameter. 3.The autism detection method based on brain function imaging data analysis of claim 2, characterized in that: The specific acquisition method of the contour movement parameter is as follows: A three-dimensional coordinate system is established according to the patient images to obtain the pixel point coordinates of each pixel point in the patient images, the patient images are processed through an image contour segmentation algorithm and an image contour tracking algorithm to obtain a patient contour, the pixel point coordinates in the patient contour are summed according to the number of pixel points in the patient contour to obtain total coordinates, the total coordinates are divided according to the number of pixel points in the patient contour to obtain a contour center point, the moving distance and the moving angle of the contour center point are calculated according to a time sequence, to obtain the moving distance and the moving angle, the moving distance and the moving angle are standardized to remove the dimension and are comprehensively calculated to obtain the contour movement parameter. 4.The autism detection method based on brain function imaging data analysis of claim 3, characterized in that: The specific acquisition method of the moving distance is as follows: In the time sequence, the distance between the contour center point of the next second and the contour center point of the last second is calculated through the Euclidean distance formula to obtain the moving distance. 5.The autism detection method based on brain function imaging data analysis of claim 2, characterized in that: The specific acquisition method of the brightness fluctuation parameter is as follows: A two-dimensional coordinate system is established according to the brain structure image, pixel point coordinates of each pixel point in the brain structure image are obtained, the brain structure image is subjected to gray scale conversion, brightness values of each pixel point are obtained, denoted as brightness, the brightness values of each pixel point in each frame of the brain structure image are summed and then averaged to obtain an equalized brightness of each frame, the brightness values of each pixel point in each frame of the brain structure image are sequentially subtracted from the equalized brightness of each frame and the absolute values are taken to obtain brightness deviation values of each pixel point, the brightness deviation values of each pixel point are summed to obtain a brightness confusion value of each frame of the brain structure image, and the brightness confusion values of each frame of the brain structure image are summed according to a time sequence to obtain a brightness fluctuation parameter. 6.The autism detection method based on brain function imaging data analysis of claim 1, characterized in that: The specific acquisition method of the brain structure abnormality factor is as follows: The brain structure image is processed by an image contour segmentation algorithm and an image contour tracking algorithm to obtain a brain contour, the brain contour of the brain structure image is comprehensively calculated to obtain a brain contour abnormality parameter, the brain contour abnormality parameter and the brightness fluctuation parameter are subjected to standardization processing to remove dimensions and are comprehensively calculated to obtain the brain structure abnormality factor; NY = LC x JC; Wherein, NY represents the brain structure abnormality factor, LC represents the brightness fluctuation parameter, and JC represents the brain contour abnormality parameter.
7. The autism detection method based on brain function imaging data analysis according to claim 6, characterized in that: The specific acquisition method of the brain contour abnormality parameter is as follows: The same kind of brain contour of different time sequences is sequentially corresponded to obtain a brain contour data set, the brain contour data set includes brain contours of various different time sequences, the area of each brain contour of each different time sequence is calculated, that is, the pixel points in each brain contour are summed to obtain the brain contour area of each different time sequence, the brain contour area of the next second and the brain contour area of the previous second of each different time sequence are subtracted and the absolute values are taken to obtain a brain contour area deviation value greater than zero, the brain contour area deviation value is summed to obtain a brain contour area deviation total value, and the brain contour area deviation total values of different kinds of brain contours are summed according to the number of kinds of brain contours to obtain the brain contour abnormality parameter. 8.The autism detection method based on brain function imaging data analysis of claim 7, characterized in that: The specific steps of sequentially corresponding the same kind of brain contour of different time sequences are as follows: Let the first brain contour in the brain structure image at the first time step be T1N1, and let a certain brain contour in the brain structure image at the second time step be T2N. j For T1N1 and T2N j The coordinates of each pixel in T1N1 are statistically analyzed. Then, the coordinates of any pixel in T2N1 are compared with those in T2N1. j The coordinates of any pixel in T1N1 are sequentially traversed and matched. If the same pixel coordinates are matched, the number is counted. If different pixel coordinates are matched, the matching continues until the pixel coordinates in T2N1 are found to be the same as those in T2N1. j All pixel coordinates in T2N have been matched, resulting in different matching similarities. A ranking algorithm is then used to sort the T2N values. j Sort the matches in ascending order of similarity, find the match with the highest similarity, and select the T2N match with that similarity. j Corresponding to T1N1, and so on, until the same brain contour in the brain structure image at the first time point corresponds to the same brain contour in the brain structure image at the second time point, and so on, the same brain contour in different time series are sequentially matched. 9.The autism detection method based on brain function imaging data analysis of claim 1, characterized in that: The specific steps of correcting the brain structure in the image by the brain structure correction algorithm are as follows: The brightness of the brain structure image is processed by adaptive histogram equalization, and finally the brightness of the brain structure image is equalized; A single frame of the corrected brain structure image is obtained by randomly selecting a single frame of the brain structure image with a brain contour abnormality parameter of zero between adjacent time sequences.
10. An autism detection system based on brain function imaging data analysis, for implementing the autism detection method based on brain function imaging data analysis of any one of claims 1-9, characterized in that, The system comprises an image data acquisition module, a historical data storage module, an image data analysis module, a data processing and correction module, and a correction success judgment module. The image data acquisition module is used for acquiring patient images and brain structure images in real time. The historical data storage module is used for storing historical normal patient images and brain structure images and calculating a judgment threshold; The image data analysis module is used for calling the image data acquisition module and the historical data storage module, and comprehensively analyzing to obtain a shaking influence factor and a brain structure abnormality factor, which are used for judging whether the patient shakes and whether the brain structure is abnormal; The data processing and correction module is used for calling the image data analysis module, issuing a prompt to the patient according to the judgment result of the patient shaking, and correcting and adjusting the brain structure image according to the judgment result of the brain structure abnormality; The correction success judgment module calls the data processing and correction module, judges whether the brain structure image is corrected successfully, and re-calls the image data analysis module according to the judgment result of the correction failure.