High myopia cataract image processing system based on image recognition

Abstract

The present application relates to the field of image processing, especially to a high myopia cataract image processing system based on image recognition, comprising a target analysis module, used to determine the image analysis strategy of the target processing object according to the image reference index and the imaging interference proportion; a reference processing module, used to determine the setting mode of the reference analysis set of the target processing image based on the feature dynamic recurrence coefficient and the recurrence area overlap coefficient, and respond to the processing reference condition to determine the reference processing mode of the target processing image; a reference division module, used to execute the setting mode of the reference analysis set of the target processing image determined by the reference processing module; a stable analysis module, used to determine whether to perform signal attenuation compensation on the target processing image according to the signal attenuation index of the execution reference image according to the effect execution difference coefficient, the present application improves the image processing efficiency of the patient's eye image.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of image processing, in particular to a high myopia cataract image processing system based on image recognition. BACKGROUND

[0002] Based on the OCT examination process, the eye examination image of the high myopia cataract patient is obtained, which is an important basis for judging the patient's symptoms and postoperative rehabilitation progress. However, there are rich interference factors in the OCT scanning process that affect the quality of the obtained OCT eye image, and the eye state of the high myopia cataract patient itself also affects the image acquisition quality. Therefore, it is often necessary to process the obtained OCT eye image to denoise and super-resolution reconstruction of the OCT image, so as to improve the effectiveness of the obtained eye image for the diagnosis process. However, the existing image processing system often fails to adjust the image processing process according to the influence of the actual patient's own state and behavior on the OCT image, resulting in low processing effect and processing efficiency of the patient's OCT image.

[0003] Chinese patent application publication No. CN118982492A discloses a jitter distortion correction image processing system for ophthalmic OCT, including a data acquisition module, using an SD-OCT acquisition device to acquire images of the human eye retina, different layers are marked with different colors in the image; a data set division module divides the collected and annotated images into different subsets to divide test data sets and training data sets; a data enhancement module performs data enhancement on the training data set; a task establishment module establishes training tasks, test tasks and detection tasks; a network model executes training tasks, test tasks and detection tasks, and uses the trained network model to process images of the human eye retina to be detected. However, the above-mentioned scheme fails to determine a targeted image processing process according to the influence of different patient's own state and behavior on the image acquisition quality, resulting in low processing efficiency of the patient's eye image. SUMMARY

[0004] Therefore, the present application provides a high myopia cataract image processing system based on image recognition to overcome the problem that the prior art fails to determine a targeted image processing process according to the influence of different patient's own state and behavior on the image acquisition quality, resulting in low processing efficiency of the patient's eye image.

[0005] To achieve the above-mentioned purpose, the present application provides a high myopia cataract image processing system based on image recognition, comprising:

[0006] The target analysis module is used to determine the image analysis strategy of the target processing object as disturbance reference analysis or detection stability analysis for the target processing object according to the image reference index and the imaging disturbance proportion.

[0007] a reference processing module, connected with the target analysis module, configured to determine a reference analysis set of the target processing image based on the feature dynamic recurrence coefficient and the recurrence region overlap coefficient, and determine a reference processing mode of the target processing image as performing phase synchronization analysis or reference matching analysis on the target processing image in response to a processing reference condition.

[0008] a reference division module, connected with the reference processing module, configured to perform the reference analysis set of the target processing image determined by the reference processing module based on the recurrence frequency stability index, or the recurrence region offset degree and the region recurrence index.

[0009] a stability analysis module, connected with the target analysis module, configured to determine whether to perform signal attenuation compensation on the target processing image according to a signal attenuation index of the execution reference image according to the effect execution difference coefficient.

[0010] Further, the image evaluation condition responded by the target analysis module is that the imaging interference index of the historical processing image is greater than the preset imaging interference index, and the historical processing image is recorded as an interference processing image.

[0011] The image reference index is determined according to the number of historical processing images of the target processing object and the acquisition evaluation coefficient.

[0012] The imaging interference proportion is the proportion of the number of interference processing images of the target processing object in the number of historical processing images of the target processing object.

[0013] Further, the target processing condition responded by the target analysis module is that the image reference index of the target processing object is greater than the preset image reference index and the imaging interference proportion is greater than the preset imaging interference proportion, and the reference processing module is determined to perform interference reference analysis on the target processing object, and the interference reference analysis includes:

[0014] determining the setting mode of the reference analysis set of the target processing image based on the feature dynamic recurrence coefficient and the recurrence region overlap coefficient, and determining the reference processing mode of the target processing image.

[0015] Further, the reference processing module responds to the reference analysis condition and detects the feature dynamic recurrence coefficient and the recurrence region overlap coefficient of the historical processing image of the target processing object.

[0016] The feature dynamic recurrence coefficient is determined based on the feature recurrence parameter of each historical processing image of the target processing object, and the feature recurrence parameter is determined according to the interval time length between the acquisition time of the abnormal recurrence scanning frame.

[0017] The recurrence area overlap coefficient is determined based on a positioning recurrence index and a feature recurrence index.

[0018] The reference analysis condition is used by the target analysis module to determine whether the reference processing module performs interference reference analysis on the target processing object in a certain processing state.

[0019] Further, the reference setting condition to which the reference processing module responds is that the feature dynamic recurrence coefficient of the target processing object is greater than a preset feature dynamic recurrence coefficient, and the reference division module determines a reference analysis set of the target processing image based on a recurrence frequency stability index.

[0020] The recurrence frequency stability index of the determined reference analysis set is greater than a preset recurrence frequency stability index.

[0021] Further, the reference setting condition to which the reference division module responds is that the feature dynamic recurrence coefficient of the target processing object is less than or equal to a preset feature dynamic recurrence coefficient and the recurrence area overlap coefficient is greater than a preset recurrence area overlap coefficient, and the reference analysis set of the target processing image is determined based on a recurrence area offset degree and an area recurrence index.

[0022] The recurrence area offset degree of any historical processing image in the determined reference analysis set is less than or equal to a preset recurrence area offset degree, and the area recurrence index of the reference analysis set is greater than a preset area recurrence index.

[0023] Further, the processing reference condition to which the reference processing module responds is that the reference analysis set is determined based on a feature similarity coefficient and a recurrence frequency stability index, and the target processing image is subjected to phase synchronization analysis.

[0024] Reference phase information of the reference analysis set is acquired to determine a reference phase difference of the target processing image, and phase correction of the target processing image is completed based on the reference phase difference.

[0025] Further, the processing reference condition to which the reference processing module responds is that the reference analysis set is determined based on a recurrence area offset degree and an area recurrence index, and the target processing image is subjected to reference matching analysis to determine a reference image of the target processing image.

[0026] The reference image is determined according to a reference matching coefficient of each continuous scanning frame of the target processing image, and the reference matching coefficient is determined according to a reference similarity coefficient and a scanning quality parameter.

[0027] Further, the target analysis module responds to the target processing condition that the image reference index of the target processing object is less than or equal to a preset image reference index or the imaging interference proportion is less than or equal to a preset imaging interference proportion, and determines that the stable analysis module performs detection stable analysis on the target processing object, and the detection stable analysis includes:

[0028] According to the execution effect correlation coefficient between each historical processing image and the target processing image, an effect execution difference coefficient is determined;

[0029] Based on the effect execution difference coefficient, it is determined whether to perform execution analysis on the target processing image.

[0030] Further, the stable analysis module responds to the analysis setting condition that the effect execution difference coefficient is less than or equal to a preset effect execution difference coefficient, and performs execution analysis on the target processing image, and the execution analysis includes:

[0031] Based on the scanning execution difference value, an execution reference image of the target processing image is determined, and the execution reference image is a historical processing image with a scanning execution difference value less than a preset scanning execution difference value;

[0032] According to the signal attenuation index of the execution reference image, signal attenuation compensation is performed on the target processing image.

[0033] Compared with the prior art, the beneficial effects of the present application are that the technical scheme of the present application determines the analysis strategy for the target processing object based on the image reference index and the imaging interference proportion of the target processing object, analyzes the processing method for the target processing image, so that the analysis process for the target processing object is more in line with the actual situation of the target processing object, and the image processing efficiency for the patient's eye image is improved.

[0034] Further, in the present application, for the target processing object with the image reference index greater than the preset image reference index and the imaging interference proportion greater than the preset imaging interference proportion, interference reference analysis is performed, the setting mode of the reference analysis set of the target processing image is determined according to the feature dynamic recurrence coefficient and the recurrence area overlap coefficient, and the processing strategy of the target processing image is determined according to the obtained reference analysis set, so as to ensure that the processing process for the target processing image is in line with the actual situation, the image abnormality rule of the target processing object is determined by analyzing the historical processing image of the target processing object, and the execution efficiency of the image processing process is improved.

[0035] Further, the setting mode of the reference analysis set of the target processing image is determined according to the feature dynamic recurrence coefficient and the recurrence area overlap coefficient in the application, the dynamic law condition and the positioning law condition of the abnormal feature of the artifact in the acquisition process of each historical processing image of the target processing object are represented based on the feature dynamic recurrence coefficient and the recurrence area overlap coefficient, and then the main factor of the abnormal artifact in the historical processing image is determined, the subsequent analysis process is preliminarily divided, and the processing efficiency of the image processing process is further improved.

[0036] Further, the reference processing mode of the target processing image is determined according to the division mode of the reference analysis set in the application, the main reference factor presented by the historical processing image in the reference analysis set is ensured to be more consistent with the processing process of the target processing image, so that the analysis efficiency in the actual processing process of the target processing image is improved while the accuracy of the processing process is ensured.

[0037] Further, the detection stability analysis is performed on the target processing object with the image reference index less than or equal to the preset image reference index or the imaging interference proportion less than or equal to the preset imaging interference proportion, whether the execution analysis is performed on the target processing image is determined based on the effect execution difference coefficient, when the effect execution difference coefficient is small, it is represented that the quality difference between the target processing image and the historical processing image is mainly caused by the scanning parameter used in the scanning process, the signal attenuation index is analyzed by matching the historical processing image with a smaller scanning parameter used in the scanning process, and the target processing image is compensated, so that the analysis efficiency in the actual processing process of the target processing image is improved while the accuracy of the processing process is ensured. BRIEF DESCRIPTION OF DRAWINGS

[0038] Fig. 1 The module connection diagram of the high myopia cataract image processing system based on image recognition in the application;

[0039] Fig. 2 The flowchart of the target analysis module for determining the target analysis strategy according to the image reference index and the imaging interference proportion in the application;

[0040] Fig. 3 The flowchart of the reference processing module for determining the setting mode of the reference analysis set based on the feature dynamic recurrence coefficient and the recurrence area overlap coefficient in the application;

[0041] Fig. 4 The flowchart of the stability analysis module for determining whether the execution analysis is performed on the target processing image according to the effect execution difference coefficient in the application. DETAILED DESCRIPTION

[0042] In order to make the objects, technical schemes and advantages of the present application clearer, the following further describes the present application with reference to the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0043] The preferred embodiments of the present application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present application and not to limit the protection scope of the present application.

[0044] It should be noted that, in the description of the present application, the terms indicating the direction or positional relationship such as "upper", "lower", "left", "right", "inner", "outer" and the like are based on the direction or positional relationship shown in the drawings, which is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application.

[0045] In addition, it should be further noted that, in the description of the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connection" should be understood in a broad sense, for example, can be fixedly connected, can be detachably connected, or integrally connected; can be mechanically connected, can be electrically connected; can be directly connected, can be indirectly connected through an intermediate medium, and can be the communication inside two elements. Those skilled in the art can understand the specific meaning of the above terms in the present application according to the specific circumstances.

[0046] Please refer to Figs. 1 to 4 As shown in the drawings, the present application provides an image recognition-based high myopia cataract image processing system, comprising:

[0047] The target analysis module is used to determine the image analysis strategy of the target processing object as interference reference analysis or detection stability analysis according to the image reference index and the imaging interference proportion;

[0048] The reference processing module is connected with the target analysis module and is used to determine the setting mode of the reference analysis set of the target processing image based on the feature dynamic recurrence coefficient and the recurrence area overlap coefficient, and in response to the processing reference condition, to determine the reference processing mode of the target processing image as phase synchronization analysis or reference matching analysis to the target processing image, so as to perform image optimization processing on the target processing image;

[0049] The reference division module is connected with the reference processing module and is used to execute the setting mode of the reference analysis set of the target processing image determined by the reference processing module, and the setting mode is to determine the reference analysis set of the target processing image based on the recurrence frequency stability index, or the recurrence area offset degree and the area recurrence index;

[0050] a stable analysis module connected with the target analysis module, configured to determine whether to perform signal attenuation compensation on the target processing image according to a signal attenuation index of the reference image according to the effect execution difference coefficient.

[0051] In the present application, the current image processing examination image is recorded as a target processing image, the target processing image corresponding patient is recorded as a target processing object, and the examination image of the target processing object after completing the image processing process is recorded as a historical processing image of the target processing object. The historical processing image and the target processing image are both eye images obtained by OCT scanning of the target patient. Since the imaging effect of the eye image of the high myopia cataract patient is poor, the processing effect of the obtained eye OCT image is of great significance to the diagnosis result of the high myopia cataract patient.

[0052] In the present application, several image processing records are applied. Each image processing record records the regional signal intensity, pixel distribution uniformity, imaging interference index, image reference index, imaging interference proportion, feature dynamic recurrence coefficient, recurrence frequency stability index, recurrence area overlap coefficient, recurrence area offset degree, area recurrence index, effect execution difference coefficient and scanning execution difference value in the processing process of the OCT examination image of the target processing object. Each image processing record corresponds to a qualified mark. The qualified mark records whether the processing efficiency of the OCT examination image of the patient meets the user's demand. It can be understood that the user can determine whether the processing efficiency of the OCT examination image of the patient meets the demand according to the self-set index. For example, the self-set index can be but is not limited to image processing efficiency, which is the maximum value of the processing time of each OCT examination image.

[0053] Specifically, the image evaluation condition responded by the target analysis module is that the imaging interference index of the historical processing image is greater than the preset imaging interference index, and the historical processing image is recorded as an interference processing image.

[0054] The image reference index is determined according to the number of historical processing images of the target processing object and the acquisition evaluation coefficient.

[0055] The imaging interference proportion is the proportion of the number of interference processing images of the target processing object in the number of historical processing images of the target processing object.

[0056] Wherein, for a single target processing object, the image reference index is the product of the number of historical processing images of the target processing object and the acquisition evaluation coefficient, the acquisition evaluation coefficient = the sum of the parameter difference values of the image acquisition parameters of each category / the number of categories of the image acquisition parameters used to determine the acquisition stability parameter, for a single category of image acquisition parameter, the parameter difference value n is the number of historical processing images of the target processing object, yi is the image acquisition parameter of the category in the acquisition process of the i-th historical processing image, y0 is the average value of the image acquisition parameter of the category in the acquisition process of each historical processing image, the parameter difference value is used to represent the variation of the image acquisition parameter used in the acquisition process of each historical processing image of the target processing object, the user can determine the category of the image acquisition parameter used to determine the acquisition stability parameter according to the actual working scene, and the category of the image acquisition parameter used to determine the acquisition stability parameter in the application includes but is not limited to: scanning density, scanning depth and scanning speed, if the imaging interference index of the historical processing image is greater than the preset imaging interference index, the historical processing image with the imaging interference index greater than the preset imaging interference index is recorded as an interference processing image, and the imaging interference proportion = the number of interference processing images of the target processing object / the number of historical processing images of the target processing object;

[0057] For a single historical processing image, the imaging interference index is the sum of the product of the image quality parameter and the disturbance region parameter and the corresponding interference evaluation coefficient, the value of the corresponding interference evaluation coefficient of the image quality parameter and the disturbance region parameter, the user can determine according to the actual working scene, for example, the user can set according to the image processing record, provide a value of the interference evaluation coefficient corresponding to the image quality parameter and the disturbance region parameter, the value of the interference evaluation coefficient corresponding to the image quality parameter is 0.4, and the value of the interference evaluation coefficient corresponding to the disturbance region parameter is 0.6;

[0058] For a single historical processing image, the image quality parameter is the sum of the lateral resolution and the axial resolution of the historical processing image, the lateral resolution refers to the minimum distance that the OCT system can distinguish two regions in the scanning direction, that is, the direction perpendicular to the light beam, the longitudinal resolution refers to the minimum distance that the OCT system can distinguish two different depth layers in the light beam direction, how to determine the lateral resolution and the axial resolution of the historical processing image is easy for those skilled in the art to understand, and will not be repeated here, the disturbance region parameter is the number of the disturbance image region of the historical processing image, the historical processing image is divided into a plurality of rectangular regions with the same area, and the divided rectangular region is recorded as a processing sub-region, if the region signal intensity of a processing sub-region is less than a preset region signal intensity or the pixel distribution uniformity is less than a preset pixel distribution uniformity, the processing sub-region is recorded as a disturbance image region, the region signal intensity is the average value of the gray values of each pixel point in the processing sub-region, the pixel distribution uniformity = the number of exceeding pixel points in the processing sub-region / the number of pixel points in the processing sub-region, the exceeding pixel point is a pixel point with a gray difference value greater than a preset gray difference value, for a single pixel point, the gray difference value is the absolute value of the difference between the gray value of the pixel point and the average value of the gray values of each pixel point in the processing sub-region, how to obtain the gray value of the pixel point in the image is easy for those skilled in the art to understand, and will not be repeated here;

[0059] The values of the preset region signal intensity, the preset gray difference value, the preset pixel distribution uniformity and the preset imaging interference index can be determined by the user according to the actual working scene, for example, the user can set according to the image processing record, the higher the user's requirement for the processing efficiency of the OCT examination image of the patient, the greater the value of the preset region signal intensity, the smaller the value of the preset gray difference value, the greater the value of the preset pixel distribution uniformity, and the smaller the value of the preset imaging interference index, a method for determining the value of the preset region signal intensity is provided, the average value of the region signal intensity of the disturbance image region in the image processing record that meets the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset region signal intensity, a method for determining the value of the preset gray difference value is provided, the value of the preset gray difference value is 40% of the gray range threshold of the processing sub-region, the gray range threshold is the difference between the maximum gray value and the minimum gray value in the processing sub-region, a method for determining the value of the preset pixel distribution uniformity is provided, the average value of the pixel distribution uniformity of the disturbance image region in the image processing record that meets the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset pixel distribution uniformity, and a method for determining the value of the preset imaging interference index is provided, the average value of the imaging interference index of the interference processing image in the image processing record that meets the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset imaging interference index.

[0060] Specifically, the target processing condition to which the target analysis module responds is that the image reference index of the target processing object is greater than the preset image reference index and the imaging interference proportion is greater than the preset imaging interference proportion, and then it is determined that the reference processing module performs interference reference analysis on the target processing object, and the interference reference analysis includes:

[0061] Based on the feature dynamic recurrence coefficient and the recurrence region overlap coefficient, the setting mode of the reference analysis set of the target processing image is determined, and the reference processing mode of the target processing image is determined.

[0062] The values of the preset image reference index and the preset imaging interference proportion can be determined by the user according to the actual working scene. For example, the user can set it according to the image processing record. The higher the user's requirement for the processing efficiency of the OCT examination image of the patient, the greater the value of the preset image reference index, and the greater the value of the preset imaging interference proportion. A method for determining the value of the preset image reference index is provided. The image processing record for interference reference analysis of the target processing object is recorded as an analysis reference record. The minimum value of the image reference index in the analysis reference record that meets the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset image reference index. A method for determining the value of the preset imaging interference proportion is provided. The minimum value of the imaging interference proportion in the analysis reference record that meets the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset imaging interference proportion.

[0063] If the image reference index of a target processing object is greater than the preset image reference index and the imaging interference proportion is greater than the preset imaging interference proportion, it indicates that the target processing object has more historical processing images for behavior analysis during the examination process, and at the same time, it contains more abnormal features that can be analyzed. Interference reference analysis is performed on the target processing object to optimize and match the targeted image processing mode, thereby improving the processing efficiency of the obtained target processing image.

[0064] Specifically, the reference processing module responds to the reference analysis condition and detects the feature dynamic recurrence coefficient and the recurrence region overlap coefficient of the historical processing image of the target processing object.

[0065] The feature dynamic recurrence coefficient is determined based on the feature recurrence parameters of each historical processing image of the target processing object, and the feature recurrence parameters are determined according to the interval time between the acquisition time of the abnormal recurrence scanning frame.

[0066] The recurrence region overlap coefficient is determined based on the positioning recurrence index and the feature recurrence index.

[0067] The reference analysis condition is used by the target analysis module to determine whether the reference processing module performs interference reference analysis on the target processing object in the first processing state.

[0068] The feature dynamic recurrence coefficient is calculated according to the following formula: f = m * f0+ (1-m) * ft m is the number of historical processing images of the target processing object, ft is the feature recurrence parameter of the tth historical processing image of the target processing image, f0 is the average value of the feature recurrence parameters of each historical processing image of the target processing image, for a single historical processing image, the feature recurrence parameter is the average value of each adjacent recurrence duration in the OCT scanning process in the process of obtaining the historical processing image, for any two abnormal recurrence scanning frames obtained in the above OCT scanning process, if there is no other abnormal recurrence scanning frame between the two abnormal recurrence scanning frames, the interval duration between the acquisition time of the two abnormal recurrence scanning frames is recorded as the adjacent recurrence duration, in the OCT scanning process, a plurality of continuous scanning frames are obtained for the target processing object, a plurality of continuous scanning frames obtained in the process of obtaining the historical processing object are divided into a plurality of rectangular regions with the same area, the number of rectangular regions obtained by different continuous scanning frames is consistent, and the rectangular regions obtained by division are recorded as scanning sub-regions, the scanning sub-regions with a region signal intensity less than a preset region signal intensity or a pixel distribution uniformity less than a preset pixel distribution uniformity are recorded as disturbance regions, and the continuous scanning frames with the disturbance regions are recorded as abnormal recurrence scanning frames.

[0069] The recurrence area overlap coefficient of the single target processing object is the sum of a positioning recurrence index and a feature recurrence index, the positioning recurrence index = the number of categories of abnormal occurrence positions existing in each historical processing image of the target processing object / the number of categories of abnormal occurrence positions corresponding to the historical processing image of the target processing object, and the feature recurrence index = the number of feature recurrence images of the target processing object / the number of historical processing images of the target processing object, for a single historical processing image, if the parameter difference value of each scan line misregistration parameter of the historical processing image is less than the preset parameter difference value corresponding to each scan line misregistration parameter, the historical processing image is determined as a feature recurrence image, for any scan misregistration parameter, the parameter difference value is the absolute value of the difference between the value of the scan misregistration parameter of the historical processing image and the average value of the values of the scan misregistration parameter of each historical processing image, the value of the preset parameter difference value can be determined by a user according to an actual working scenario, for example, the user can set the value according to image processing records, the higher the requirement of the user on the processing efficiency of the OCT image of the patient, the smaller the value of the preset parameter difference value, a value of the preset parameter difference value is provided, the value of the preset parameter difference value is 5% of the average value of the values of the scan misregistration parameter of each historical processing image, and the scan line misregistration parameters used for determining the feature recurrence image in the application include but are not limited to misregistration amplitude and misregistration frequency.

[0070] Specifically, the reference setting condition to which the reference processing module responds is that the feature dynamic recurrence coefficient of the target processing object is greater than a preset feature dynamic recurrence coefficient, and then it is determined that the reference division module determines the reference analysis set of the target processing image based on the recurrence frequency stability index.

[0071] The recurrence frequency stability index of the determined reference analysis set is greater than a preset recurrence frequency stability index.

[0072] The value of the preset feature dynamic recurrence coefficient can be determined by a user according to an actual working scenario, for example, the user can set the value according to image processing records, the higher the requirement of the user on the processing efficiency of the OCT image of the patient, the greater the value of the preset feature dynamic recurrence coefficient, a value of the preset feature dynamic recurrence coefficient is provided, a method for determining the value of the preset feature dynamic recurrence coefficient is provided, image processing records in which the reference analysis set of the target processing image is determined based on the recurrence frequency stability index are recorded as first division reference records, and the minimum value of the feature dynamic recurrence coefficient of the target processing object in the first division reference records meeting the requirement of the user on the processing efficiency of the OCT image of the patient is recorded as the preset feature dynamic recurrence coefficient.

[0073] The reference analysis set is a set of several historical processing images of the target processing object. For the determined reference analysis set, the recurrence frequency stability index = average value of the recurrence frequency parameters of each historical processing image in the reference analysis set / difference between the maximum and minimum values of the recurrence frequency parameters of each historical processing image in the reference analysis set. For a single historical processing image, the recurrence frequency parameter is the number of abnormal recurrence scanning frames obtained per unit time in the OCT scanning process of the historical processing image. The user can determine the value of the preset recurrence frequency stability index according to the actual working scene. For example, the user can set it according to the image processing record. The higher the user's requirement for the processing efficiency of the OCT examination image of the patient, the greater the value of the preset recurrence frequency stability index. A method for determining the value of the preset recurrence frequency stability index is provided. The minimum value of the recurrence frequency stability index of the reference analysis set in the first division reference record that meets the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset recurrence frequency stability index.

[0074] Specifically, the reference setting condition to which the reference processing module responds is that the feature dynamic recurrence coefficient of the target processing object is less than or equal to a preset feature dynamic recurrence coefficient and the recurrence region overlap coefficient is greater than a preset recurrence region overlap coefficient, and then it is determined that the reference division module determines the reference analysis set of the target processing image based on the recurrence region offset degree and the region recurrence index.

[0075] The recurrence region offset degree of any historical processing image in the determined reference analysis set is less than or equal to a preset recurrence region offset degree and the region recurrence index of the reference analysis set is greater than a preset region recurrence index.

[0076] The value of the preset recurrence region overlap coefficient can be determined by the user according to the actual working scene. For example, the user can set it according to the image processing record. The higher the user's requirement for the processing efficiency of the OCT examination image of the patient, the greater the value of the preset recurrence region overlap coefficient. A method for determining the value of the preset recurrence region overlap coefficient is provided. The image processing record of the reference analysis set of the target processing image determined based on the recurrence region offset degree and the region recurrence index is recorded as the second division reference record. The minimum value of the recurrence region overlap coefficient of the target processing object in the second division reference record that meets the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset recurrence region overlap coefficient.

[0077] The region recurrence index is the number of historical processing images in the reference analysis set that have the same abnormality occurrence position as the target processing image, the abnormality occurrence position is the specific position in the eye region of the target processing object corresponding to the disturbance region in the continuous scanning frame during the process of obtaining the historical processing image or the target processing image, the categories of the abnormality occurrence position include but are not limited to: choroid region, retinal blood vessel region, optic disc surrounding blood vessel dense area and macular region, for any historical processing image in the reference analysis set, the recurrence region offset degree = the number of offset recurrence scanning frames of the historical processing image / the number of abnormality recurrence scanning frames of the historical processing image, for a single abnormality occurrence position corresponding to the disturbance region in the abnormality recurrence scanning frame of the historical processing image, if the number of associated scanning frames of the abnormality occurrence position is the maximum value of the number of associated scanning frames of each abnormality occurrence position corresponding to the historical processing object, the associated scanning frame is recorded as the offset recurrence scanning frame, the associated scanning frame is the abnormality recurrence scanning frame in which the disturbance region corresponds to the abnormality occurrence position;

[0078] The user can determine the values of the preset recurrence region offset degree and the preset region recurrence index according to the actual working scene, for example, the user can set according to the image processing record, the higher the user's requirement for the processing efficiency of the OCT examination image of the patient, the smaller the value of the preset recurrence region offset degree, the larger the value of the preset region recurrence index, a method for determining the value of the preset recurrence region offset degree is provided, the average value of the recurrence region offset degrees of each historical processing image in the reference analysis set in the second divided reference record that meets the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset recurrence region offset degree, and a method for determining the value of the preset recurrence frequency stability index is provided, the minimum value of the region recurrence indices of the reference analysis set in the second divided reference record that meets the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset region recurrence index.

[0079] Specifically, the processing reference condition to which the reference processing module responds is that the reference analysis set is determined based on the feature similarity coefficient and the recurrence frequency stability index, then the phase synchronization analysis is performed on the target processing image;

[0080] The reference phase information of the reference analysis set is obtained to determine the reference phase difference of the target processing image, and the phase correction of the target processing image is completed based on the reference phase difference.

[0081] Specifically, the reference processing module responds to the processing reference condition that the reference analysis set determines based on the recurrence area offset degree and the area recurrence index, and performs reference matching analysis on the target processing image based on the reference matching parameters and the scanning quality parameters to determine the reference image of the target processing image.

[0082] The reference image is determined according to the reference matching coefficients of each continuous scanning frame of the target processing image, and the reference matching coefficients are determined according to the reference similarity coefficients and the scanning quality parameters.

[0083] If the reference analysis set of the target processing image is determined based on the feature similarity coefficients and the recurrence frequency index, it indicates that the historical processing images of the target processing object are mainly inclined to be the reference factors of the processing process of the target processing image due to the dynamic presentation law caused by the breathing or heartbeat of the target processing object and other factors. The image abnormalities caused by such factors often have periodic rules. By determining the phase conditions of the artifact regions in each historical processing image in the reference analysis set, and performing phase correction on the target processing image based on the difference between the phase information of the target processing image and the reference analysis set, the accuracy of the phase correction is ensured while the analysis efficiency of the phase correction process is improved. The same abnormal occurrence positions of the disturbance regions in the continuous scanning frames of the target processing image are analyzed for the abnormal recurrence scanning frames, and the abnormal recurrence scanning frames with the same abnormal occurrence positions of the disturbance regions in the continuous scanning frames of the target processing image are recorded as phase analysis frames. The phase information of each pixel point in the phase analysis frame is recorded as reference phase information. For a single abnormal recurrence scanning frame of the target processing image, the phase analysis frame with the largest image similarity coefficient of the abnormal recurrence scanning frame is recorded as the phase synchronization frame of the abnormal recurrence scanning frame. The phase difference between the pixel points of the disturbance regions of the abnormal recurrence scanning frame and its phase synchronization frame is calculated. The average value of the phase difference between the pixel points of the disturbance regions is recorded as the reference phase difference. If the reference phase difference is greater than a preset reference phase difference, the phase correction is performed on the abnormal recurrence scanning frame, and the processing mode of the phase correction process is determined based on the change direction of the phase difference. The processing mode includes but is not limited to translation correction and rotation processing. How to calculate the phase difference between the pixel points and how to perform phase correction are mastered by those skilled in the art, and are not described here.

[0084] The preset reference phase difference value can be determined by the user according to the actual working scene, for example, the user can set it according to the image processing record, the higher the user's requirement for the processing efficiency of the OCT examination image of the patient, the smaller the preset reference phase difference value, and a method for determining the preset reference phase difference value is provided, and the minimum value of the reference phase difference of the abnormal repeated scanning frame when performing phase correction in the image processing record meeting the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset reference phase difference;

[0085] If the reference analysis set of the target processing image is determined based on the repetition area offset degree and the area repetition index, it indicates that the historical processing image of the target processing object is mainly inclined to the image abnormality caused by the eye movement of the target processing object as the reference factor of the processing process of the target processing image, the determined reference image has a great influence on the image processing result in the process of image processing for the abnormal situation caused by the eye movement of the target processing object, the reference image determined in the historical processing image processing process in the reference analysis set provides reference for the determination of the reference image of the target processing image, while ensuring the reliability of the reference image determination result, the analysis efficiency of the reference image determination process is improved, therefore, the reference image of the target processing image is determined by analyzing the similarity between each continuous scanning frame of the target processing image and the reference image of each historical processing image in the reference analysis set and the scanning quality of each continuous scanning frame, the continuous scanning frame with the maximum reference matching coefficient is recorded as the reference image of the target processing image, for a single continuous scanning frame of the target processing image, the reference matching coefficient = ln (the reference similarity coefficient of the continuous scanning frame x the scanning quality parameter of the continuous scanning frame), the scanning quality parameter is the sum of the image signal-to-noise ratio and the blood vessel density of the continuous scanning frame, how to determine the image signal-to-noise ratio and the blood vessel density of the continuous scanning frame is easy to understand for those skilled in the art, and will not be described here, the reference similarity coefficient is the average value of the image similarity coefficient between the continuous scanning frame and the reference image of each historical processing image in the reference analysis set, for the continuous scanning frame and any reference image, the user can determine the image similarity coefficient between the continuous scanning frame and the reference image according to the feature point matching method, the feature point matching process includes: obtaining the feature points of the continuous scanning frame and the reference image, and generating high-order vectors of each feature point, calculating the Euclidean distance between the high-order vectors of each feature point in the continuous scanning frame and the reference image, if the Euclidean distance between the high-order vectors of two feature points is less than a preset matching distance, the two feature points are recorded as a matching pair, and the image similarity coefficient is the number of matching pairs determined;

[0086] The preset matching distance value can be determined by the user according to the actual working scene, for example, the user can set it according to the image processing record, the higher the user's requirement for the processing efficiency of the OCT examination image of the patient, the smaller the preset matching distance value, a method for determining the preset matching distance value is provided, the average value of the Euclidean distance between the high-order vectors in each matching pair in the image processing record meeting the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset matching distance, the eye movement of other continuous scanning frames is judged by the obtained reference image, so as to eliminate the abnormal features in the target processing image, how to determine the feature points and their high-order vectors, how to determine the Euclidean distance and the matching pair, and how to complete the subsequent image processing according to the reference image are easy for those skilled in the art to understand, and will not be described here.

[0087] Specifically, the target analysis module responds to the target processing condition that the image reference index of the target processing object is less than or equal to the preset image reference index or the imaging interference proportion is less than or equal to the preset imaging interference proportion, and determines that the stable analysis module detects the target processing object for stable analysis, and the stable analysis includes:

[0088] According to the execution effect correlation coefficient between each historical processing image and the target processing image, an effect execution difference coefficient is determined.

[0089] Based on the effect execution difference coefficient, it is determined whether to perform analysis on the target processing image.

[0090] Among them, for a single historical processing image, the execution effect correlation coefficient between the historical processing image and the target processing image is the absolute value of the difference between the imaging interference indexes of the historical processing image and the target processing image / the scanning execution difference value between the historical processing image and the target processing image, and the scanning execution difference value is the sum of the execution difference degrees of each scanning execution parameter, for a single scanning execution parameter, the execution difference degree is the absolute value of the difference between the numerical values of the scanning execution parameter in the scanning process of the historical processing image and the target processing image / the numerical value of the scanning execution parameter in the scanning process of the target processing image, and the categories of scanning execution parameters used to determine the scanning execution difference value in the application include but are not limited to: light source bandwidth, objective numerical aperture, imaging depth, A-scan rate and B-scan density;

[0091] The effect execution difference coefficient m is the number of historical processing images of the target processing object, Xj is the execution effect correlation coefficient of the jth historical processing image and the target processing image, y0 is the average value of the execution effect correlation coefficients of each historical processing image and the target processing image, and the value of the preset effect execution difference coefficient can be determined by the user according to the actual working scene. For example, the user can set it according to the image processing record. The higher the user's requirement for the processing efficiency of the OCT examination image of the patient, the larger the value of the preset effect execution difference coefficient. A value determination method of the preset effect execution difference coefficient is provided. The image processing record subjected to execution analysis for the target processing image is recorded as a difference reference record. The minimum value of the effect execution difference coefficient in the difference reference record meeting the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset effect execution difference coefficient.

[0092] Specifically, the analysis setting condition to which the stability analysis module responds is that the effect execution difference coefficient is less than or equal to the preset effect execution difference coefficient, and then execution analysis is performed on the target processing image. The execution analysis includes:

[0093] Based on the scanning execution difference value, the execution reference image of the target processing image is determined. The execution reference image is a historical processing image with a scanning execution difference value less than the preset scanning execution difference value.

[0094] According to the signal attenuation index of the execution reference image, signal attenuation compensation is performed on the target processing image.

[0095] If the effect execution difference coefficient is less than or equal to the preset effect execution difference coefficient, execution analysis is performed on the target processing image. If the effect execution difference coefficient is greater than the preset effect execution difference coefficient, execution analysis is not performed on the target processing image. The value of the preset scanning execution difference value can be determined by the user according to the actual working scene. For example, the user can set it according to the image processing record. The higher the user's requirement for the processing efficiency of the OCT examination image of the patient, the smaller the value of the preset scanning execution difference value. A value determination method of the preset scanning execution difference value is provided. The average value of the scanning execution difference values of the execution reference images in the image processing record meeting the user's requirement for the processing efficiency of the OCT examination image of the patient is recorded as the preset scanning execution difference value.

[0096] The process of determining the signal attenuation index and performing signal attenuation compensation on the target processing image includes: performing depth profile analysis on each execution reference image, measuring the signal intensity at different depths, and using an exponential attenuation model to fit the change of the signal intensity with the depth. The exponential attenuation model used in the present application is I(Z) = I0e -aZWherein a is the signal attenuation index, which reflects the attenuation characteristics of the optical signal in biological tissue, I0 is the initial signal intensity, and I(Z) is the signal intensity when the depth or signal propagation distance is Z. According to the determined signal attenuation index, the attenuation degree of the signal at different depths or signal propagation distances in the image acquisition process is determined, and the signal intensity at different depths of the target processing image is compensated, thereby obtaining the target processing image with signal attenuation compensation. This is a content that those skilled in the art have mastered, and will not be described here.

[0097] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without deviating from the principles of the present application, and the technical solutions after these changes or replacements will fall within the protection scope of the present application.

[0098] The above description is only the preferred embodiments of the present application and is not intended to limit the present application; for those skilled in the art, the present application can have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included within the protection scope of the present application.

Claims

1. An image recognition-based high myopia cataract image processing system, characterized in that, The method comprises the following steps: a target analysis module is configured to determine an image analysis strategy for a target processing object according to an image reference index and an imaging interference proportion, wherein the image analysis strategy is interference reference analysis or detection stability analysis for the target processing object; a reference processing module is connected to the target analysis module and is configured to determine a setting mode of a reference analysis set of the target processing image based on a feature dynamic recurrence coefficient and a recurrence area overlap coefficient, and determine a reference processing mode of the target processing image according to a processing reference condition, wherein the reference processing mode is phase synchronization analysis or benchmark matching analysis for the target processing image, and the target processing image is subjected to image optimization processing; a reference division module is connected to the reference processing module and is configured to execute the setting mode of the reference analysis set of the target processing image determined by the reference processing module, wherein the setting mode is determined based on a recurrence frequency stability index, a recurrence area offset degree, and a region recurrence index; a stability analysis module is connected to the target analysis module and is configured to determine whether to perform signal attenuation compensation on the target processing image according to an execution reference image signal attenuation index according to an execution difference coefficient; For a single target processing object, the image reference index is the product of the number of historical processing images of the target processing object and the acquisition evaluation coefficient, the acquisition evaluation coefficient = the sum of parameter difference values of image acquisition parameters of each category / the number of categories of image acquisition parameters used to determine the acquisition stability parameter, for a single category of image acquisition parameter, the parameter difference value , n is the number of historical processing images of the target processing object, yi is the image acquisition parameter of the i-th historical processing image in the acquisition process, y0 is the average value of the image acquisition parameter of each historical processing image in the acquisition process, the imaging interference proportion is the proportion of the number of interference processing images of the target processing object in the number of historical processing images of the target processing object, the feature dynamic recurrence coefficient , m is the number of historical processing images of the target processing object, ft is the feature recurrence parameter of the t-th historical processing image of the target processing image, f0 is the average value of the feature recurrence parameter of each historical processing image of the target processing image, the recurrence region overlap coefficient is the sum of the positioning recurrence index and the feature recurrence index, the positioning recurrence index = the number of categories of abnormal occurrence positions existing in each historical processing image of the target processing object / the number of categories of abnormal occurrence positions corresponding to the historical processing images of the target processing object, the feature recurrence index = the number of feature recurrence images of the target processing object / the number of historical processing images of the target processing object; for the determined reference analysis set, the recurrence frequency stability index is an average value of recurrence frequency parameters of each historical processing image in the reference analysis set divided by a difference between a maximum value and a minimum value of the recurrence frequency parameters of each historical processing image in the reference analysis set, for a single historical processing image, the recurrence frequency parameter is the number of abnormal recurrence scanning frames obtained per unit time in the OCT scanning process of the historical processing image, for any historical processing image in the reference analysis set, the recurrence area offset degree is the number of offset recurrence scanning frames of the historical processing image divided by the number of abnormal recurrence scanning frames of the historical processing image, and the region recurrence index is the number of historical processing images in the reference analysis set having the same abnormal occurrence position as the target processing image.

2. The image recognition based high myopia cataract image processing system according to claim 1, wherein, The image evaluation condition to which the target analysis module responds is that the imaging interference index of a historical processing image is greater than a preset imaging interference index, and the historical processing image is recorded as an interference processing image; The image reference index is determined according to the number of historical processing images of the target processing object and a collection evaluation coefficient.

3. The image recognition based high myopia cataract image processing system according to claim 2, characterized in that, The target processing condition to which the target analysis module responds is that the image reference index of the target processing object is greater than a preset image reference index and the imaging interference proportion is greater than a preset imaging interference proportion, and the reference processing module is determined to perform interference reference analysis on the target processing object, wherein the interference reference analysis comprises: determining a setting mode of a reference analysis set of the target processing image based on a feature dynamic recurrence coefficient and a recurrence area overlap coefficient, and determining a reference processing mode of the target processing image.

4. The image recognition based high myopia cataract image processing system according to claim 3, characterized in that, The reference processing module responds to the reference analysis condition and detects the feature dynamic recurrence coefficient and the recurrence area overlap coefficient of the historical processing image of the target processing object; The feature dynamic recurrence coefficient is determined based on a feature recurrence parameter of each historical processing image of the target processing object, and the feature recurrence parameter is determined according to the interval time between the acquisition time of abnormal recurrence scanning frames. The recurrence region overlap coefficient is determined based on a positioning recurrence index and a feature recurrence index. The reference analysis condition is used by the target analysis module to determine whether the reference processing module performs interference reference analysis on the target processing object in a certain processing state.

5. The image recognition based high myopia cataract image processing system according to claim 4, characterized in that, The reference setting condition to which the reference processing module responds is that the feature dynamic recurrence coefficient of the target processing object is greater than a preset feature dynamic recurrence coefficient, and the reference division module determines a reference analysis set of the target processing image based on a recurrence frequency stability index. The recurrence frequency stability index of the determined reference analysis set is greater than a preset recurrence frequency stability index.

6. The image recognition based high myopia cataract image processing system according to claim 5, wherein, The reference setting condition to which the reference division module responds is that the feature dynamic recurrence coefficient of the target processing object is less than or equal to a preset feature dynamic recurrence coefficient and the recurrence region overlap coefficient is greater than a preset recurrence region overlap coefficient, and the reference division module determines a reference analysis set of the target processing image based on a recurrence region offset and a region recurrence index. The recurrence region offset of any historical processing image in the determined reference analysis set is less than or equal to a preset recurrence region offset, and the region recurrence index of the reference analysis set is greater than a preset region recurrence index.

7. The image recognition based high myopia cataract image processing system according to claim 6, characterized in that, The processing reference condition to which the reference processing module responds is that the reference analysis set is determined based on a feature similarity coefficient and a recurrence frequency stability index, and the reference processing module performs phase synchronization analysis on the target processing image. Reference phase information of the reference analysis set is obtained to determine a reference phase difference of the target processing image, and phase correction of the target processing image is completed based on the reference phase difference.

8. The image recognition based high myopia cataract image processing system according to claim 7, characterized in that, The processing reference condition to which the reference processing module responds is that the reference analysis set is determined based on a recurrence region offset and a region recurrence index, and the reference processing module performs reference matching analysis on the target processing image to determine a reference image of the target processing image. The reference image is determined according to a reference matching coefficient of each continuous scanning frame of the target processing image, and the reference matching coefficient is determined according to a reference similarity coefficient and a scanning quality parameter.

9. The image recognition based high myopia cataract image processing system according to claim 8, wherein, The target processing condition to which the target analysis module responds is that the image reference index of the target processing object is less than or equal to a preset image reference index or the imaging interference proportion is less than or equal to a preset imaging interference proportion, and the stable analysis module determines to perform detection stability analysis on the target processing object, and the detection stability analysis includes: An effect execution difference coefficient is determined according to an execution effect correlation coefficient between each historical processing image and the target processing image. Whether to perform execution analysis on the target processing image is determined based on the effect execution difference coefficient.

10. The image recognition based high myopia cataract image processing system according to claim 9, wherein, The analysis setting condition to which the stable analysis module responds is that the effect execution difference coefficient is less than or equal to a preset effect execution difference coefficient, and the execution analysis on the target processing image is performed, and the execution analysis includes: An execution reference image of the target processing image is determined based on a scanning execution difference value, and the execution reference image is a historical processing image with a scanning execution difference value less than a preset scanning execution difference value. Signal attenuation compensation is performed on the target processing image according to a signal attenuation index of the execution reference image.

Images