Eye image intelligent management method and system based on artificial intelligence, and medium

Abstract

The invention discloses an intelligent eye image management method and system based on artificial intelligence and a medium, and relates to the technical field of medical image data management. Eye image data are divided into common eye image data and key eye image data; the relevance between the eye image data is analyzed, and the storage mode of the eye image data is adjusted; storing the common eye image data after lossy compression and the key eye image data after lossless compression; the associated eye image data are stored together, so that the query efficiency of personnel can be improved; the eye image data are subjected to lossy compression, and the eye image data subjected to lossy compression and lossless compression are stored together, so that the rendering speed can be accelerated, and the performance pressure can be reduced; when the person obtains the inquired eye image data, the nearby eye image is also loaded, so that the rendering speed of the eye image data can be further accelerated.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image data management, and specifically to an intelligent management method, system and medium for eye images based on artificial intelligence. Background Art

[0002] Eye image data covers a variety of imaging technologies, such as fundus photography, optical coherence tomography, scanning laser ophthalmoscopy, fundus angiography, visual field examination, etc. Different imaging technologies can provide eye information at different levels and structures, and have characteristics such as high resolution and rich details; ophthalmologists can use eye image data to study the pathogenesis and pathophysiological processes of eye diseases, and by analyzing a large number of patient image data, find early signs and characteristic changes of diseases, which helps to deeply understand the nature of diseases and helps ophthalmologists train their skills; however, precisely because of the characteristics of high resolution and rich details, the loading and rendering speed of eye image data will generate high performance pressure, affecting the reading efficiency of eye image data; therefore, how to improve the reading efficiency of eye image data has become an urgent problem to be solved. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent management method, system and medium for eye images based on artificial intelligence to solve the problems raised in the prior art.

[0004] To achieve the above purpose, the present invention provides the following technical solution: An intelligent management system for eye images based on artificial intelligence, including an eye image acquisition module, a data storage module, an eye image data management module and an interaction module; the output end of the eye image acquisition module is connected to the input end of the data storage module, and eye image data is obtained through an ophthalmic imaging device; the data storage module is connected to the eye image data management module and is used to store eye image data; the eye image data management module optimizes the storage method of eye image data based on the access data of personnel to eye image data; the interaction module is connected to the data storage module and is used to provide eye image data to personnel.

[0005] Specifically, the eye image data management module further includes a storage optimization unit, a correlation analysis unit, an eye image data classification unit and a significance test unit; the correlation analysis unit is used to analyze the correlation between eye image data; the eye image data classification unit is used to classify eye image data into ordinary eye image data and key eye image data; the significance test unit is used to perform a significance test on the data; the storage optimization unit is used to optimize the storage method of eye image data.

[0006] Specifically, the data storage module further includes a compression unit, a transmission unit, and a sorting unit; the compression unit is used to perform lossy compression and lossless compression on ordinary eye image data, and perform lossless compression on key eye image data; the transmission unit is used to transmit the eye image data to the interaction module; the sorting unit optimizes the storage order of the lossy compressed ordinary eye image data and the lossless compressed key eye image data based on the optimization result of the eye image data management module.

[0007] To achieve the above object, the present invention provides the following technical solutions: An intelligent management method for eye images based on artificial intelligence, including the following steps:

[0008] Obtain eye image data, and divide the eye image data into ordinary eye image data and key eye image data according to the information contained in the eye image data;

[0009] Analyze the correlation between eye image data, and adjust the storage method of eye image data; perform lossy compression and lossless compression on ordinary eye image data, and perform lossless compression on key eye image data; store the lossy compressed ordinary eye image data and the lossless compressed key eye image data, and establish a link between the lossy compressed ordinary eye image data and the lossless compressed ordinary eye image data.

[0010] Specifically, the step of dividing the eye image data into ordinary eye image data and key eye image data according to the information contained in the eye image data further includes the following steps:

[0011] Obtain the characteristics of the eye image data, including the usage frequency, the single usage time, and the score of the contained information; the score of the contained information is determined by the key area situation contained in the eye image data; use the characteristics of the eye image data to train a binary classification neural network model; extract a part of the eye image data as the training data set, and the unextracted eye image data as the verification set, and assign two different labels to the eye image data in the training data set, corresponding to lossy compression and lossless compression respectively; use the characteristics of the eye image data as the input and the label as the output to train the binary classification neural network model; use the trained binary classification neural network model to classify the eye image data in the verification set, and judge whether the classification result meets the requirements. If so, the binary classification neural network model meets the requirements, otherwise, re-extract the eye image data to train the binary classification neural network model.

[0012] Specifically, the step of analyzing the correlation between eye image data further includes the following steps:

[0013] Let x i represent the i-th eye image data, and i represents the label of the eye image data; obtain the person's eye image data xi Access records to determine the eye image data x i and x j The correlation P between ij , x j represents the j-th eye image data, and i represents the label of the eye image data; P ij = N(i, j) / N(i), where N(i) represents the number of sets containing the eye image data x i The number of sets, and N(i, j) represents the number of sets that contain both the eye image data x i and x j ; Using the access records of the person to the eye image data x i , simultaneously obtain the access order of the person to the eye image data x i and x j ; In the set that contains both the eye image data x i and x j , obtain the number of times n1 that the person accesses the eye image data x i first and the number of times n2 that the person accesses the eye image data x i later. Conduct a significance test on n1 to determine whether there is a significant difference between n1 and 0. If there is a significant difference, then mark the eye image data x i before x j ; If there is no significant difference between n1 and 0, then conduct a significance test on n2 to determine whether there is a significant difference between n2 and 0. If there is a significant difference between n2 and 0, then mark the eye image data x i after x j ; If there is no significant difference between n2 and 0 either, then there is no order marking for the eye image data x i and x j .

[0014] Specifically, storing the lossy-compressed ordinary eye image data and the lossless-compressed key eye image data further includes the following steps:

[0015] Set the initial temperature, set the initial storage order of the eye image data, and determine the loss value corresponding to the initial storage order; Use the initial temperature as the temperature index T;

[0016] Step 1, for variables n = 1, 2,..., H, repeat Step 2; H is the set number of loops;

[0017] Step 2, optimize the storage order of the eye image data by generating perturbations on the basis of the current storage order;

[0018] Step 3: Lower the temperature index T according to the temperature reduction plan. If the temperature index T is greater than or equal to the set threshold, return to Step 1; if the temperature index T is less than the set threshold, obtain the storage order of the eye image data according to the current solution.

[0019] Specifically, the optimization of the storage order of the eye image data by generating perturbations on the basis of the current storage order specifically includes the following steps:

[0020] S10: Represent the storage order of the current eye image data as x1, x2, …, xm, where x1, x2, …, xm represent the eye image data; the order of other eye image data x2, …, xm except x1 is not determined; for k = 2, 3, …, m, execute Step S20 to complete one round of perturbation.

[0021] S20: Obtain the correlation between the eye image data xk and the eye image data whose order is not determined, sort the eye image data whose order is not determined in descending order to obtain a new storage order; determine the loss value corresponding to the new storage order; calculate the loss value increment brought by the new storage order. If the increment is less than 0, accept the new storage order; if the increment is greater than or equal to 0, accept the new storage order with a certain probability; determine the storage order of the eye image data before and including xk according to the accepted storage order.

[0022] Specifically, the loss value is determined through the following steps:

[0023] S100: Based on the historical access information of the personnel to the eye image data, determine the access frequencies of the personnel to different sets containing the eye image data, and splice the access frequencies of the user to different sets containing the eye image data into a closed interval. Each set containing the eye image data corresponds to a section on the closed interval. Randomly generate num numbers, where num is a positive integer, and determine the set containing the eye image data required by the personnel according to the interval where the random number is located.

[0024] S200: For each set containing the eye image data determined in Step S100, judge the order marks between the eye image data in the set. If there are no order marks between the eye image data in the set, obtain the sum of the distances between the eye image data in the set under the storage order of the eye image data as the loss value; the distance is the number of eye image data outside the set contained between the eye image data in the set under the storage order of the eye image data.

[0025] If there is an order mark among the eye image data in the set, first obtain the loss value without the order mark. Let xa and xb represent two eye image data whose storage order of the eye image data is different from the order mark among the eye image data in the set. Obtain the distance between xa and xb in the storage order of the eye image data. If the distance exceeds the threshold, increment the additional search count by one. If the distance does not exceed the threshold, keep the additional search count unchanged. The threshold is set according to the rendering and loading performance of the search device. Determine the product of the additional search count and a reference value, and add the product to the loss value without the order mark to obtain the loss value with the order mark.

[0026] To achieve the above object, the present invention provides the following technical solution: An intelligent management medium for eye images based on artificial intelligence, wherein a computer program is stored in the medium. It is characterized in that when the computer program is executed by a processor, the steps of an intelligent management method for eye images based on artificial intelligence are implemented.

[0027] Compared with the prior art, the beneficial effects of the present invention are as follows: Storing related eye image data together can improve the query efficiency of personnel; performing lossy compression on the eye image data and storing the lossy-compressed and lossless-compressed eye image data together can accelerate the rendering speed and reduce the performance pressure; when personnel obtain the queried eye image data, nearby eye images will also be loaded, which can further accelerate the rendering speed of the eye image data; finally, it can reduce the search times of personnel and relieve the performance pressure on the data storage module. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic structural diagram of an intelligent management system for eye images based on artificial intelligence of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0030] Embodiment: As Figure 1As shown in the figure, the present invention provides a technical solution, an intelligent management system for eye images based on artificial intelligence, including an eye image acquisition module, a data storage module, an eye image data management module, and an interaction module; the output end of the eye image acquisition module is connected to the input end of the data storage module, and eye image data is obtained through an ophthalmic imaging device; the data storage module is interconnected with the eye image data management module and is used to store eye image data; the eye image data management module optimizes the storage method of eye image data based on the access data of personnel to the eye image data; the interaction module is interconnected with the data storage module and is used to provide eye image data to personnel.

[0031] The eye image data management module further includes a storage optimization unit, a correlation analysis unit, an eye image data classification unit, and a significance test unit; the correlation analysis unit is used to analyze the correlation between eye image data; the eye image data classification unit is used to classify eye image data into ordinary eye image data and key eye image data; the significance test unit is used to perform a significance test on the data; the storage optimization unit is used to optimize the storage method of eye image data.

[0032] The data storage module further includes a compression unit, a transmission unit, and a sorting unit; the compression unit is used to perform lossy compression and lossless compression on ordinary eye image data and perform lossless compression on key eye image data; the transmission unit is used to transmit eye image data to the interaction module; the sorting unit optimizes the storage order of lossy compressed ordinary eye image data and lossless compressed key eye image data based on the optimization result of the eye image data management module.

[0033] Embodiment: The present invention provides a technical solution, an intelligent management medium for eye images based on artificial intelligence, wherein a computer program is stored in the medium, and characterized in that when the computer program is executed by a processor, the steps of an intelligent management method for eye images based on artificial intelligence are implemented.

[0034] Embodiment: The present invention provides a technical solution, an intelligent management method for eye images based on artificial intelligence, including the following steps:

[0035] Obtain eye image data, and classify the eye image data into ordinary eye image data and key eye image data according to the information contained in the eye image data;

[0036] Analyze the correlation between eye image data, and adjust the storage method of eye image data; perform lossy compression and lossless compression on ordinary eye image data, and perform lossless compression on key eye image data; store the lossy compressed ordinary eye image data and the lossless compressed key eye image data, and establish a link between the lossy compressed ordinary eye image data and the lossless compressed ordinary eye image data.

[0037] Obtain the characteristics of eye image data, including usage frequency, single usage time, and score of contained information; the score of contained information is determined by the situation of key regions contained in the eye image data; train a binary classification neural network model using the characteristics of eye image data; extract a part of the eye image data as the training dataset, and the unextracted eye image data as the validation set, and assign two different labels to the eye image data in the training dataset, corresponding to lossy compression and lossless compression respectively; use the characteristics of the eye image data as the input and the label as the output to train the binary classification neural network model; use the trained binary classification neural network model to classify the eye image data in the validation set, and determine whether the classification result meets the requirements. If so, the binary classification neural network model meets the requirements, otherwise, re-extract the eye image data to train the binary classification neural network model.

[0038] The situation of key regions contained can be determined according to whether it contains lesion regions such as microaneurysms, bleeding points, and macula. Different scores are assigned to the lesion regions such as microaneurysms, bleeding points, and macula, and the scores of all in the eye image data are added up to obtain the score of contained information. Extract a part of the eye image data as the training data, and manually assign labels to the extracted training data. If the extracted eye image data does not contain key information and allows the image quality to be reduced, then assign the label 1, corresponding to lossy compression; if the extracted eye image data contains key information and does not allow the image quality to be reduced, then assign the label 0, corresponding to lossless compression; after the binary classification neural network model is trained, extract a part of the eye image data in the validation set for verification, and obtain the first error rate and the second error rate according to the verification situation. Among them, the first error rate is the probability that the output result of the eye image data that does not allow the image quality to be reduced in the binary classification neural network model is 1; the second error rate is the probability that the output result of the eye image data that allows the image quality to be reduced in the binary classification neural network model is 0. The first and second error rates are weighted to obtain the index of the binary classification neural network model; judge whether the binary classification neural network model meets the requirements according to the index.

[0039] The analysis of the correlation between eye image data further includes the following steps:

[0040] Let x iRepresents the i-th eye image data, where i represents the label of the eye image data; obtain the access record of the person to the eye image data x i and determine the correlation P i between the eye image data x j and x ij , x j represents the j-th eye image data, and i represents the label of the eye image data; P ij = N(i, j) / N(i), where N(i) represents the number of sets containing the eye image data x i , and N(i, j) represents the number of sets that contain both the eye image data x i and x j ; use the access record of the person to the eye image data x i to simultaneously obtain the access order of the person to the eye image data x i and x j . In the set that contains both the eye image data x i and x j , obtain the number of times n1 that the person accesses the eye image data x i first and the number of times n2 that the person accesses the eye image data x i later. Conduct a significance test on n1 to determine whether there is a significant difference between n1 and 0. If there is a significant difference, then mark the eye image data x i before x j ; if there is no significant difference between n1 and 0, then conduct a significance test on n2 to determine whether there is a significant difference between n2 and 0. If there is a significant difference between n2 and 0, then mark the eye image data x i after x j . If there is also no significant difference between n2 and 0, then there is no order mark for the eye image data x i and x j .

[0041] When an ophthalmologist obtains eye image data for research, learning, or skill training, multiple related or associated eye image data are usually obtained at one time, and these multiple eye image data form a set; based on the set that contains both the eye image data x i and x j , the correlation between the eye image data x i and x j can be obtained; based on the access order of the ophthalmologist to the eye image data x i and x j , determine whether the ophthalmologist has a habitual access pattern to the eye image data, such as a tendency to obtain x i first and then obtain x j ; determine x iand x j The purpose of the order of j is to make the arrangement of the eye image data conform to the habits of the person, thereby reducing the adverse effects caused by the habitual operations of the person; for example, for the eye image data x i If it is always obtained last, then when the person obtains the eye image data x i At this time, it is easy to habitually think that all data has been obtained, even if there are other eye image data after the eye image data x i exists.

[0042] To determine whether there is a significant difference between n1 and 0, at a significance level of 0.05, if it meets the significance level, there is a 95% probability that it can be explained that x i was obtained before x j If there is no significant difference between n1, n2 and 0, then x i and x j There is no specific access order between them.

[0043] The storage of the lossy compressed ordinary eye image data and the lossless compressed key eye image data further includes the following steps:

[0044] Set the initial temperature, set the initial storage order of the eye image data, and determine the loss value corresponding to the initial storage order; use the initial temperature as the temperature index T;

[0045] Step 1, for variables n = 1, 2,..., H, repeat Step 2; H is the set number of loops;

[0046] Step 2, optimize the storage order of the eye image data by generating perturbations on the basis of the current storage order;

[0047] Step 3, reduce the temperature index T according to the temperature reduction scheme. If the temperature index T is greater than or equal to the set threshold, return to Step 1; if the temperature index T is less than the set threshold, obtain the storage order of the eye image data according to the current solution.

[0048] The optimization of the storage order of the eye image data by generating perturbations on the basis of the current storage order specifically includes the following steps:

[0049] S10, represent the current storage order of the eye image data with x1, x2,..., xm, where x1, x2,..., xm represent the eye image data; the order of other eye image data x2,..., xm except x1 is not determined; for k = 2, 3,..., m, execute Step S20 to complete one round of perturbation;

[0050] S20. Obtain the correlation between the eye image data xk and the eye image data whose order is not determined, sort the eye image data whose order is not determined in descending order to obtain a new storage order; determine the loss value corresponding to the new storage order; calculate the loss value increment brought by the new storage order. If the increment is less than 0, accept the new storage order. If the increment is greater than or equal to 0, accept the new storage order with a certain probability; according to the accepted storage order, determine the storage order of the eye image data before and including xk.

[0051] The loss value is determined through the following steps:

[0052] S100. Based on the historical access information of the person to the eye image data, determine the access frequency of the person to different sets containing the eye image data, and splice the access frequencies of the user to different sets containing the eye image data into a closed interval. Each set containing the eye image data corresponds to a section on the closed interval. Randomly generate num numbers, where num is a positive integer, and determine the set containing the eye image data required by the person according to the interval where the random number is located;

[0053] S200. For each set containing the eye image data determined in step S100, judge the order marks between the eye image data in the set. If there are no order marks between the eye image data in the set, obtain the sum of the distances between the eye image data in the set under the storage order of the eye image data as the loss value; the distance is the number of eye image data outside the set included between the eye image data in the set under the storage order of the eye image data.

[0054] If there are order marks between the eye image data in the set, first obtain the loss value when there are no order marks. Let xa and xb represent two eye image data whose storage order of the eye image data is different from the order marks between the eye image data in the set; obtain the distance between xa and xb under the storage order of the eye image data. If the distance exceeds the threshold, increment the additional search times by one. If the distance does not exceed the threshold, keep the additional search times unchanged; the threshold is set according to the rendering and loading performance of the search device; determine the product of the additional search times and a reference value, and add the product to the loss value when there are no order marks to obtain the loss value when there are order marks.

[0055] For example, if there are different sets of eye image data for a person, a closed interval of [0, 5] can be generated. When the access frequency of each set is the same, [0, 1], [1, 2], …, [4, 5], etc. respectively correspond to a set. When a random number falls within the interval, the loss value is calculated using the set corresponding to the interval; after repeating num times, the average value of the loss values for the num times is taken to obtain the final loss value. The reference value can be adjusted according to the situation; if there are few sequential tags between the eye image data and most of the eye image data do not have sequential tags, the reference value can be decreased; conversely, it can be increased; when there is a sequential mark between xa and xb and xa is before xb, if xa is after xb in the storage order of the eye image data, then when the user finds xb, they may subconsciously think that xa already exists and thus ignore xa, which may ultimately lead to an increase in the number of searches by the person.

[0056] When a person queries eye image data through a device such as a computer, in addition to the queried eye image data, the eye image data near the queried eye image data will also be rendered and loaded. If the nearby eye image data is not the data needed by the person, it will waste computer resources and affect the person's query efficiency;

[0057] If the nearby eye image data is also the data needed by the person, it can improve the person's query efficiency; in addition, when the person obtains the queried eye image data, the nearby eye images will also be loaded, which can improve the rendering speed of the eye image data; finally, it can reduce the number of searches by the person and relieve the performance pressure on the data storage module.

[0058] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be regarded as limiting the claimed rights.

Claims

1. An intelligent management method for eye images based on artificial intelligence, characterized in that, It includes the following steps: Obtain eye image data, and divide the eye image data into ordinary eye image data and key eye image data according to the information contained in the eye image data; Analyze the correlation between the eye image data, and adjust the storage method of the eye image data; perform lossy compression and lossless compression on the ordinary eye image data, and perform lossless compression on the key eye image data; store the lossy compressed ordinary eye image data and the lossless compressed key eye image data, and establish a link between the lossy compressed ordinary eye image data and the lossless compressed ordinary eye image data.

2. The intelligent management method for eye images based on artificial intelligence according to claim 1, wherein The step of dividing the eye image data into ordinary eye image data and key eye image data according to the information contained in the eye image data further includes the following steps: Obtain the characteristics of the eye image data, including the usage frequency, single usage time, and the score of the contained information; the score of the contained information is determined by the key area situation contained in the eye image data; train a binary classification neural network model using the characteristics of the eye image data; extract a part of the eye image data as the training data set, and the unextracted eye image data as the validation set, and assign two different labels to the eye image data in the training data set, corresponding to lossy compression and lossless compression respectively; use the characteristics of the eye image data as the input and the label as the output to train the binary classification neural network model; use the trained binary classification neural network model to classify the eye image data in the validation set, and judge whether the classification result meets the requirements. If so, the binary classification neural network model meets the requirements, otherwise re-extract the eye image data to train the binary classification neural network model.

3. The intelligent management method for eye images based on artificial intelligence according to claim 2, characterized in that, The step of analyzing the correlation between the eye image data further includes the following steps: Let x i represents the i-th eye image data, i represents the label of the eye image data; obtain the person's eye image data x i Access records to determine eye image data x i and x j The correlation between ij , x j represents the jth eye image data, i represents the label of the eye image data; P ij =N(i, j) / N(i), where N(i) represents the eye image data x i The number of sets, N(i, j) indicates that it contains eye image data x i and x j The number of sets; using personnel's eye image data x i Access records of personnel and obtain eye image data x i and x j The access order of the eye image data x i and x j In the collection, the acquisition personnel first access the eye image data x i The number of times n1 and the subsequent access to the eye image data x i The number of times n2 is used to verify the significance of n1 and determine whether there is a significant difference between n1 and 0. If there is a significant difference, the eye image data x i Marked with x j Before; if there is no significant difference between n1 and 0, then perform significance verification on n2 to determine whether there is a significant difference between n2 and 0. If there is a significant difference between n2 and 0, then convert the eye image data x i Marked with x j Then, if there is no significant difference between n2 and 0, the eye image data x i and x j No sequence mark.

4. The intelligent management method for eye images based on artificial intelligence according to claim 3, wherein, The step of storing the lossy compressed ordinary eye image data and the lossless compressed key eye image data further includes the following steps: Set the initial temperature, set the initial storage order of the eye image data, and determine the loss value corresponding to the initial storage order; use the initial temperature as the temperature index T; Step 1, for variables n = 1, 2, …, H, repeat Step 2; H is the set number of cycles; Step 2, optimize the storage order of the eye image data by generating perturbations on the basis of the current storage order; Step 3, reduce the temperature index T according to the temperature reduction scheme. If the temperature index T is greater than or equal to the set threshold, return to Step 1; if the temperature index T is less than the set threshold, obtain the storage order of the eye image data according to the current solution.

5. The intelligent management method for eye images based on artificial intelligence according to claim 4, wherein The step of optimizing the storage order of the eye image data by generating perturbations on the basis of the current storage order specifically includes the following steps: S10, represent the current storage order of the eye image data as x1, x2, …, xm, where x1, x2, …, xm represent the eye image data; the order of the other eye image data x2, …, xm except x1 is not determined; for k = 2, 3, …, m, execute Step S20 to complete one round of perturbation; S20. Obtain the correlation between the eye image data xk and the eye image data whose order has not been determined, sort the eye image data whose order has not been determined in descending order to obtain a new storage order; determine the loss value corresponding to the new storage order; calculate the loss value increment brought by the new storage order. If the increment is less than 0, accept the new storage order. If the increment is greater than or equal to 0, accept the new storage order with a certain probability; according to the accepted storage order, determine the storage order of the eye image data before and including xk.

6. The intelligent management method for eye images based on artificial intelligence according to claim 5, characterized in that The loss value is determined through the following steps: S100. Based on the historical access information of the personnel to the eye image data, determine the access frequencies of the personnel to different sets containing the eye image data, and splice the access frequencies of the user to different sets containing the eye image data into a closed interval. Each set containing the eye image data corresponds to a segment on the closed interval. Randomly generate num numbers, where num is a positive integer, and determine the set containing the eye image data required by the personnel according to the interval where the random numbers are located. S200. For each set containing the eye image data determined in step S100, judge the order marks between the eye image data in the set. If there are no order marks between the eye image data in the set, obtain the sum of the distances between the eye image data in the set under the storage order of the eye image data as the loss value; the distance is the number of eye image data outside the set between the eye image data in the set under the storage order of the eye image data. If there are order marks between the eye image data in the set, first obtain the loss value when there are no order marks. Let xa and xb represent two eye image data whose storage order of the eye image data is different from the order marks between the eye image data in the set; obtain the distance between xa and xb under the storage order of the eye image data. If the distance exceeds the threshold, increment the additional search times by one. If the distance does not exceed the threshold, keep the additional search times unchanged; the threshold is set according to the rendering and loading performance of the search device; determine the product of the additional search times and the reference value, and add the product to the loss value when there are no order marks to obtain the loss value when there are order marks.

7. An intelligent management system for eye images based on artificial intelligence, characterized in that, It includes an eye image acquisition module, a data storage module, an eye image data management module, and an interaction module; the output end of the eye image acquisition module is connected to the input end of the data storage module, and the eye image data is obtained through an ophthalmic imaging device; the data storage module is connected to the eye image data management module and is used to store the eye image data; the eye image data management module optimizes the storage method of the eye image data based on the access data of the personnel to the eye image data; the interaction module is connected to the data storage module and is used to provide the eye image data to the personnel.

8. The intelligent management system for eye images based on artificial intelligence according to claim 7, wherein, The eye image data management module further includes a storage optimization unit, a correlation analysis unit, an eye image data classification unit, and a significance test unit; the correlation analysis unit is used to analyze the correlation between eye image data; the eye image data classification unit is used to classify eye image data into ordinary eye image data and key eye image data; the significance test unit is used to perform a significance test between data; the storage optimization unit is used to optimize the storage method of eye image data.

9. The intelligent management system for eye images based on artificial intelligence according to claim 7, characterized in that, The data storage module further includes a compression unit, a transmission unit, and a sorting unit; the compression unit is used to perform lossy compression and lossless compression on ordinary eye image data and perform lossless compression on key eye image data; the transmission unit is used to transmit eye image data to the interaction module; the sorting unit optimizes the storage order of lossy compressed ordinary eye image data and lossless compressed key eye image data based on the optimization result of the eye image data management module.

10. An intelligent management medium for eye images based on artificial intelligence, wherein a computer program is stored in the medium, characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images