Depression identification method and system based on eye movement image identification

Abstract

The invention belongs to the technical field of image analysis, and particularly discloses a depression recognition method and system based on eye movement image recognition, and the method comprises the steps: carrying out the corresponding image processing and closed edge detection through collecting an eye monitoring image frame of a target object, determining an eye edge contour and a pupil image region, and then carrying out the feature extraction, the method comprises the steps of obtaining a feature value sequence and a reference center coordinate, constructing a feature matrix based on the features of each image frame, and finally performing automatic identification and classification of a classification model by using the feature matrix to obtain a classification result, so that a psychologist can perform efficient depression judgment on a target object according to the classification result. Through intelligent eye image feature extraction and eye action classification recognition, depression judgment basis based on object eye feature recognition can be efficiently provided for psychologists, so that the psychologists can quickly and accurately judge the depression condition of the target object, and important support is provided for subsequent psychological intervention treatment.

Description

Technical Field

[0001] The present invention belongs to the technical field of image analysis, and in particular relates to a depression recognition method and system based on eye movement image recognition. Background Art

[0002] Depression, also known as depressive disorder, is a mental disorder with a high incidence, high clinical cure rate, but low treatment acceptance rate and high recurrence rate. Its main feature is significant and persistent low mood, which may even be accompanied by psychotic symptoms such as delusions and hallucinations. When conducting depression examinations, psychologists mainly conduct face-to-face interviews and observe the target object's demeanor, expression and behavior to complete psychological assessments. Among them, changes in demeanor are mainly reflected in changes in eye movements and states, which are more difficult to detect. At present, the method of capturing the target object's eye movements and state changes through manual observation is highly dependent on the personal experience of psychologists, and is prone to errors and misjudgments. A more accurate and effective analysis and identification method is needed to assist psychologists in efficient depression diagnosis. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for identifying depression based on eye movement image recognition, so as to solve the above-mentioned problems existing in the prior art.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions:

[0005] In a first aspect, the present invention provides a method for identifying depression based on eye movement image recognition, comprising:

[0006] Obtain an eye monitoring video segment of the target object in a set test scene, and extract a number of sequentially arranged eye monitoring image frames from the eye monitoring video segment at intervals;

[0007] Performing image grayscale processing on the eye monitoring image frame to obtain a corresponding eye grayscale image, and performing image filtering processing on the eye grayscale image to obtain a corresponding eye filtered grayscale image;

[0008] Performing image binarization processing on the eye filtered grayscale image to obtain a corresponding eye binary image;

[0009] Performing closed edge detection on the eye binary image to determine the eye edge contour, and determining the eye corner points on both sides of the eye edge contour and the pupil image area within the eye edge contour;

[0010] Connect the eye corner points on both sides to obtain a horizontal reference line, and select a number of reference points at equal intervals on the horizontal reference line, and construct a vertical reference line on each reference point, so that the vertical reference line is perpendicular to the horizontal reference line, and the two ends of the vertical reference line respectively abut the eye edge contour;

[0011] Taking the eye corner point on one side as the origin and the horizontal reference line as the horizontal coordinate, constructing the image coordinate axis, and determining the coordinates of the center point of the pupil image area, and taking the coordinates of the center point of the pupil image area as the reference center coordinates of the corresponding eye monitoring image frame;

[0012] The ratio of the longitudinal reference line to the transverse reference line of each reference point is used as its characteristic value, and the characteristic values ​​of each reference point are arranged in sequence to obtain a characteristic value sequence of the corresponding eye monitoring image frame;

[0013] constructing a first feature sequence based on a feature value sequence of each eye monitoring image frame, constructing a second feature sequence based on a reference center coordinate of each eye monitoring image frame, and constructing a feature matrix based on the first feature sequence and the second feature sequence;

[0014] The feature matrix is ​​input into a preset eye movement classification model for classification and recognition, and the corresponding eye movement classification results are obtained. The eye movement classification results are displayed to the user, so that the user can identify and determine depression of the target object based on the eye movement classification results.

[0015] In a possible design, the grayscale processing of the eye monitoring image frame to obtain a corresponding eye grayscale image, and the image filtering processing of the eye grayscale image to obtain a corresponding eye filtered grayscale image include:

[0016] Determine the RGB three-color channel component value of each pixel in the eye monitoring image frame, and perform weighted summation of the RGB three-color channel component value of each pixel with a set weight ratio to obtain the grayscale value of each pixel, and obtain the eye grayscale image based on the grayscale value of each pixel;

[0017] The eye grayscale image is processed by Gaussian filtering algorithm to obtain the corresponding eye filtered grayscale image.

[0018] In a possible design, performing image binarization processing on the eye filtered grayscale image to obtain a corresponding eye binary image includes:

[0019] The Bernsen algorithm is used to perform image binarization on the eye filtered grayscale image to obtain the corresponding eye binary image.

[0020] In a possible design, the closed edge detection is performed on the eye binary image to determine the eye edge contour, and the eye corner points on both sides of the eye edge contour and the pupil image area within the eye edge contour, including:

[0021] The Canny edge detection operator is used to perform edge detection on the eye binary image, and the contours of each closed edge are determined based on the edge detection results;

[0022] A closed edge contour with an area within a set first area range and corner points on both sides is used as an eye edge contour, and the corner points on both sides of the eye edge contour are used as eye corner points on both sides;

[0023] A closed edge contour whose area within the eye edge contour is within a set second area range is used as a pupil edge contour, and an image area within the pupil edge contour is used as a pupil image area.

[0024] In a possible design, constructing a first feature sequence based on a feature value sequence of each eye monitoring image frame includes:

[0025] Taking the average of the feature values ​​in the same order in the feature value sequence of each eye monitoring image frame to obtain the feature average value of the corresponding order;

[0026] Combine the average values ​​of each feature in order according to the corresponding ranking to obtain the first feature sequence {Z1, Z2, ..., Z n}, where Z n The average value of the features ranked by n.

[0027] In a possible design, constructing a second feature sequence based on the reference center coordinates of each eye monitoring image frame includes:

[0028] The reference center coordinates of each eye monitoring image frame are combined according to the order of the corresponding eye monitoring image frame to obtain a second feature sequence {(X1, Y1), (X2, Y2), ..., (X m , Y m )}, where X m and Y m Respectively represent the reference center horizontal coordinate and the reference center vertical coordinate ranked as m.

[0029] In a possible design, constructing a feature matrix based on the first feature sequence and the second feature sequence includes:

[0030] Using the first feature sequence {Z1, Z2, ..., Z n} and the second characteristic sequence {(X1, Y1), (X2, Y2), ..., (X m , Y m )}Construct feature matrix Among them, m=n.

[0031] In one possible design, the eye movement classification model adopts a trained ResNet model.

[0032] In a second aspect, the present invention provides a depression identification system based on eye movement image recognition, comprising an image acquisition unit, a graying filter unit, a binary processing unit, an edge detection unit, a feature calibration unit, a coordinate determination unit, a feature arrangement unit, a matrix construction unit and a classification identification unit, wherein:

[0033] An image acquisition unit, used to acquire an eye monitoring video segment of a target object in a set test scene, and extract a plurality of sequentially arranged eye monitoring image frames from the eye monitoring video segment at intervals;

[0034] A graying filter unit is used to perform image graying processing on the eye monitoring image frame to obtain a corresponding eye gray image, and perform image filtering processing on the eye gray image to obtain a corresponding eye filtered gray image;

[0035] A binary processing unit, used for performing image binarization processing on the eye filtered grayscale image to obtain a corresponding eye binary image;

[0036] An edge detection unit is used to perform closed edge detection on the eye binary image, determine the eye edge contour, and determine the eye corner points on both sides of the eye edge contour and the pupil image area within the eye edge contour;

[0037] A feature calibration unit is used to connect the eye corner points on both sides to obtain a horizontal reference line, and select a number of reference points at equal intervals on the horizontal reference line, and construct a vertical reference line on each reference point, so that the vertical reference line is perpendicular to the horizontal reference line, and the two ends of the vertical reference line are respectively in contact with the eye edge contour;

[0038] A coordinate determination unit, used to construct an image coordinate axis with one of the eye corner points as the origin and the horizontal reference line as the horizontal coordinate, and determine the coordinates of the center point of the pupil image area, and use the coordinates of the center point of the pupil image area as the reference center coordinates of the corresponding eye monitoring image frame;

[0039] A feature arrangement unit, used to take the ratio of the longitudinal reference line to the transverse reference line of each reference point as its feature value, and arrange the feature values ​​of each reference point in sequence to obtain a feature value sequence of the corresponding eye monitoring image frame;

[0040] a matrix construction unit, configured to construct a first feature sequence based on a feature value sequence of each eye monitoring image frame, construct a second feature sequence based on a reference center coordinate of each eye monitoring image frame, and construct a feature matrix based on the first feature sequence and the second feature sequence;

[0041] The classification and recognition unit is used to input the feature matrix into a preset eye movement classification model for classification and recognition, obtain the corresponding eye movement classification results, and display the eye movement classification results to the user, so that the user can identify and judge the depression of the target object based on the eye movement classification results.

[0042] In a third aspect, the present invention provides a depression identification system based on eye movement image recognition, comprising:

[0043] A memory for storing instructions;

[0044] A processor is used to read the instructions stored in the memory and execute any one of the methods described in the first aspect according to the instructions.

[0045] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein instructions are stored on the computer-readable storage medium, and when the instructions are executed on a computer, the computer executes any one of the methods described in the first aspect. At the same time, a computer program product is also provided, and when the computer program product is executed on a computer, the computer executes any one of the methods described in the first aspect.

[0046] Beneficial effects: The present invention collects the eye monitoring image frames of the target object, performs corresponding image processing and closed edge detection, determines the eye edge contour and pupil image area, then performs feature extraction to obtain the feature value sequence and reference center coordinates, and then constructs a feature matrix based on the features of each image frame. Finally, the feature matrix is ​​used to automatically identify and classify the classification model to obtain the classification result, so that the psychologist can efficiently determine the depression of the target object based on the classification result. The present invention can efficiently provide psychologists with a basis for judging depression based on the eye feature recognition of the object through intelligent eye image feature extraction and eye movement classification and recognition, so that psychologists can quickly and accurately determine the depression condition of the target object, providing important support for subsequent psychological intervention treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] Figure 1 A schematic diagram of the method steps provided in Example 1 of the present invention;

[0049] Figure 2 A schematic diagram of the system structure provided in Example 2 of the present invention;

[0050] Figure 3 This is a schematic diagram of the system structure provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0051] It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention. The specific structure and function details disclosed herein are only used to describe the exemplary embodiments of the present invention. However, the present invention can be embodied in many alternative forms, and it should not be understood that the present invention is limited to the embodiments set forth herein.

[0052] It should be understood that, unless otherwise clearly specified and limited, the term "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the embodiments can be understood according to specific circumstances.

[0053] In the following description, certain details are provided to facilitate a complete understanding of the example embodiments. However, it will be appreciated by those of ordinary skill in the art that the example embodiments may be implemented without these certain details. For example, the system may be shown in a block diagram to avoid obscuring the example with unnecessary details. In other embodiments, well-known processes, structures, and techniques may not be shown in unnecessary detail to avoid obscuring the embodiments.

[0054] Embodiment 1:

[0055] This embodiment provides a method for identifying depression based on eye movement image recognition, which can be applied to corresponding data processing terminals, such as Figure 1 As shown, the method comprises the following steps:

[0056] S1. Obtain an eye monitoring video segment of the target object in a set test scene, and extract a number of sequentially arranged eye monitoring image frames from the eye monitoring video segment at intervals.

[0057] In specific implementation, in a set test scenario, such as when a psychologist conducts a psychological interview with a target subject, the target subject's eye monitoring video can be collected through a video collector, and then the target subject's eye monitoring video can be edited to obtain the target subject's eye monitoring video segment, and then the target subject's eye monitoring video segment can be transmitted to a data processing terminal. The data processing terminal extracts a number of sequentially arranged eye monitoring image frames from the eye monitoring video segment at a set time interval for subsequent processing.

[0058] S2. Perform image grayscale processing on the eye monitoring image frame to obtain a corresponding eye grayscale image, and perform image filtering processing on the eye grayscale image to obtain a corresponding eye filtered grayscale image.

[0059] In specific implementation, the data processing terminal first determines the RGB three-color channel component value of each pixel in the eye monitoring image frame, and performs weighted summation of the RGB three-color channel component value of each pixel with a set weight ratio to obtain the grayscale value of each pixel; then, an eye grayscale image is obtained based on the grayscale value of each pixel; then, a Gaussian filtering algorithm is used to perform Gaussian filtering on the eye grayscale image to filter out Gaussian noise in the eye grayscale image to obtain a corresponding eye filtered grayscale image.

[0060] S3. Perform image binarization processing on the eye filtered grayscale image to obtain a corresponding eye binary image.

[0061] In specific implementation, the data processing terminal can use the Bernsen algorithm to perform image binarization processing on the eye filter grayscale image to obtain the corresponding eye binary image. The Bernsen algorithm is an adaptive local threshold binarization method, which determines the threshold by calculating the maximum and minimum values ​​in the neighborhood around each pixel in the image, thereby realizing the image binarization processing. That is, two thresholds S and TT are preset, and for each pixel in the image, the maximum value M and the minimum value N in the neighborhood window centered on the pixel are calculated; if the difference between M and N is greater than S, the threshold T of the pixel is set to the average value of M and N; if the difference is less than or equal to S, the binarization result of the pixel is determined according to the relationship between T and TT. This algorithm can adaptively adjust the threshold according to the local characteristics of the eye filter grayscale image, thereby improving the binarization effect.

[0062] S4. Perform closed edge detection on the eye binary image to determine the eye edge contour, and determine the eye corner points on both sides of the eye edge contour and the pupil image area within the eye edge contour.

[0063] In specific implementation, the data processing terminal can use the Canny edge detection operator to perform edge detection on the eye binary image, and determine each closed edge contour according to the edge detection result; then use the closed edge contour with an area within a set first area range and corner points on both sides as the eye edge contour, and use the corner points on both sides of the eye edge contour as the eye corner points on both sides; then use the closed edge contour within the eye edge contour with an area within a set second area range as the pupil edge contour, and use the image area within the pupil edge contour as the pupil image area.

[0064] S5. Connect the eye corner points on both sides to obtain a horizontal reference line, and select a number of reference points at equal intervals on the horizontal reference line, and construct a vertical reference line on each reference point, so that the vertical reference line is perpendicular to the horizontal reference line, and the two ends of the vertical reference line respectively touch the eye edge contour.

[0065] In specific implementation, the data processing terminal obtains a horizontal reference line by connecting the eye corner points on both sides, and then selects a number of reference points at equal intervals on the horizontal reference line, and the number of reference points is consistent with the number of eye monitoring image frames. Then, a vertical reference line is constructed on each reference point, so that the vertical reference line is perpendicular to the horizontal reference line, and the two ends of the vertical reference line respectively abut the edge contour of the eye.

[0066] S6. Take one of the eye corner points as the origin and the horizontal reference line as the horizontal coordinate to construct the image coordinate axis, and determine the center point coordinates of the pupil image area, and use the center point coordinates of the pupil image area as the reference center coordinates of the corresponding eye monitoring image frame.

[0067] In specific implementation, the data processing terminal can use one of the eye corners (such as the left eye corner) as the origin and the horizontal reference line as the horizontal coordinate to construct the image coordinate axis. Then, based on the image coordinate axis, the center point coordinates of the pupil image area are determined, and the center point coordinates of the pupil image area are used as the reference center coordinates of the corresponding eye monitoring image frame.

[0068] S7. Taking the ratio of the longitudinal reference line to the transverse reference line of each reference point as its characteristic value, and arranging the characteristic values ​​of each reference point in sequence to obtain a characteristic value sequence of the corresponding eye monitoring image frame.

[0069] In the specific implementation, the data processing terminal first calculates the ratio of the longitudinal reference line to the transverse reference line of each reference point, and then uses the ratio of the longitudinal reference line to the transverse reference line as the characteristic value of the corresponding reference point, and then arranges and combines the characteristic values ​​of each reference point in the order of each reference point to obtain the characteristic value sequence of the corresponding eye monitoring image frame.

[0070] S8. Construct a first feature sequence based on the feature value sequence of each eye monitoring image frame, construct a second feature sequence based on the reference center coordinates of each eye monitoring image frame, and construct a feature matrix based on the first feature sequence and the second feature sequence.

[0071] In specific implementation, the data processing terminal averages the feature values ​​in the same order in the feature value sequence of each eye monitoring image frame to obtain the feature average value of the corresponding order. Then, the feature average values ​​are sequentially combined according to the corresponding order to obtain the first feature sequence {Z1, Z2, ..., Z n}, where Z n The average value of the features ranked by n.

[0072] The reference center coordinates of each eye monitoring image frame are combined according to the order of the corresponding eye monitoring image frame to obtain a second feature sequence {(X1, Y1), (X2, Y2), ..., (X m , Y m )}, where X m and Ym Respectively represent the reference center horizontal coordinate and the reference center vertical coordinate ranked as m.

[0073] Finally, using the first feature sequence {Z1, Z2, ..., Z n} and the second characteristic sequence {(X1, Y1), (X2, Y2), ..., (X m , Y m )}Construct feature matrix Among them, m=n.

[0074] S9. Input the feature matrix into a preset eye movement classification model for classification and identification, obtain the corresponding eye movement classification results, and display the eye movement classification results to the user, so that the user can identify and determine depression of the target object based on the eye movement classification results.

[0075] In specific implementation, the data processing terminal can input the feature matrix into a preset eye movement classification model for classification and identification, and obtain the corresponding eye movement classification result. The eye movement classification model uses a trained ResNet model, and the eye movement classification result is the classification identification label of the target object's eye movement, which represents the eye movement type of the target object in the set scene. The data processing terminal finally displays the eye movement classification result to the user, so that the user can identify and determine depression of the target object based on the eye movement classification result.

[0076] The method of this embodiment can efficiently provide psychologists with a basis for judging depression based on the eye feature recognition of the target object through intelligent eye image feature extraction and eye movement classification and recognition, so that psychologists can quickly and accurately judge the depression condition of the target object, providing important support for subsequent psychological intervention treatment.

[0077] Embodiment 2:

[0078] This embodiment provides a depression recognition system based on eye movement image recognition, such as Figure 2 As shown, it includes an image acquisition unit, a graying filter unit, a binary processing unit, an edge detection unit, a feature calibration unit, a coordinate determination unit, a feature arrangement unit, a matrix construction unit and a classification recognition unit, wherein:

[0079] An image acquisition unit, used to acquire an eye monitoring video segment of a target object in a set test scene, and extract a plurality of sequentially arranged eye monitoring image frames from the eye monitoring video segment at intervals;

[0080] A graying filter unit is used to perform image graying processing on the eye monitoring image frame to obtain a corresponding eye gray image, and perform image filtering processing on the eye gray image to obtain a corresponding eye filtered gray image;

[0081] A binary processing unit, used for performing image binarization processing on the eye filtered grayscale image to obtain a corresponding eye binary image;

[0082] An edge detection unit is used to perform closed edge detection on the eye binary image, determine the eye edge contour, and determine the eye corner points on both sides of the eye edge contour and the pupil image area within the eye edge contour;

[0083] A feature calibration unit is used to connect the eye corner points on both sides to obtain a horizontal reference line, and select a number of reference points at equal intervals on the horizontal reference line, and construct a vertical reference line on each reference point, so that the vertical reference line is perpendicular to the horizontal reference line, and the two ends of the vertical reference line are respectively in contact with the eye edge contour;

[0084] A coordinate determination unit, used to construct an image coordinate axis with one of the eye corner points as the origin and the horizontal reference line as the horizontal coordinate, and determine the coordinates of the center point of the pupil image area, and use the coordinates of the center point of the pupil image area as the reference center coordinates of the corresponding eye monitoring image frame;

[0085] A feature arrangement unit, used to take the ratio of the longitudinal reference line to the transverse reference line of each reference point as its feature value, and arrange the feature values ​​of each reference point in sequence to obtain a feature value sequence of the corresponding eye monitoring image frame;

[0086] a matrix construction unit, configured to construct a first feature sequence based on a feature value sequence of each eye monitoring image frame, construct a second feature sequence based on a reference center coordinate of each eye monitoring image frame, and construct a feature matrix based on the first feature sequence and the second feature sequence;

[0087] The classification and recognition unit is used to input the feature matrix into a preset eye movement classification model for classification and recognition, obtain the corresponding eye movement classification results, and display the eye movement classification results to the user, so that the user can identify and judge the depression of the target object based on the eye movement classification results.

[0088] Embodiment 3:

[0089] This embodiment provides a depression recognition system based on eye movement image recognition, such as Figure 3 As shown, including:

[0090] Data interface, used to establish data connection between the processor and the external data terminal;

[0091] A memory for storing execution instructions;

[0092] The processor is used to read the execution instruction stored in the memory, and execute the depression identification method based on eye movement image recognition in Example 1 according to the execution instruction.

[0093] Optionally, the system further includes an internal bus, through which the processor, the memory and the data interface can be interconnected, and the internal bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0094] The memory may include, but is not limited to, random access memory (RAM), read only memory (ROM), flash memory, first input first output (FIFO) and / or first in last out (FILO) memory, etc. The processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0095] Embodiment 4:

[0096] This embodiment provides a computer-readable storage medium, on which execution instructions are stored. When the execution instructions are executed on a computer, the computer executes the depression identification method based on eye movement image recognition in Embodiment 1. The computer-readable storage medium refers to a carrier for storing data, which may include but is not limited to a floppy disk, an optical disk, a hard disk, a flash memory, a USB flash drive and / or a memory stick, etc. The computer may be a general-purpose computer, a special-purpose computer, a computer network or other programmable devices.

[0097] This embodiment also provides a computer program product, which, when executed on a computer, executes the depression identification method based on eye movement image recognition in Embodiment 1. The computer may be a general-purpose computer, a special-purpose computer, a computer network or other programmable device.

[0098] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for identifying depression based on eye movement image recognition, characterized in that: include: Obtain an eye monitoring video segment of the target object in a set test scene, and extract a number of sequentially arranged eye monitoring image frames from the eye monitoring video segment at intervals; Performing image grayscale processing on the eye monitoring image frame to obtain a corresponding eye grayscale image, and performing image filtering processing on the eye grayscale image to obtain a corresponding eye filtered grayscale image; Performing image binarization processing on the eye filtered grayscale image to obtain a corresponding eye binary image; Performing closed edge detection on the eye binary image to determine the eye edge contour, and determining the eye corner points on both sides of the eye edge contour and the pupil image area within the eye edge contour; Connect the eye corner points on both sides to obtain a horizontal reference line, and select a number of reference points at equal intervals on the horizontal reference line, and construct a vertical reference line on each reference point, so that the vertical reference line is perpendicular to the horizontal reference line, and the two ends of the vertical reference line respectively abut the eye edge contour; Taking the eye corner point on one side as the origin and the horizontal reference line as the horizontal coordinate, constructing the image coordinate axis, and determining the coordinates of the center point of the pupil image area, and taking the coordinates of the center point of the pupil image area as the reference center coordinates of the corresponding eye monitoring image frame; The ratio of the longitudinal reference line to the transverse reference line of each reference point is used as its characteristic value, and the characteristic values ​​of each reference point are arranged in sequence to obtain a characteristic value sequence of the corresponding eye monitoring image frame; constructing a first feature sequence based on a feature value sequence of each eye monitoring image frame, constructing a second feature sequence based on a reference center coordinate of each eye monitoring image frame, and constructing a feature matrix based on the first feature sequence and the second feature sequence; The feature matrix is ​​input into a preset eye movement classification model for classification and recognition, and the corresponding eye movement classification results are obtained. The eye movement classification results are displayed to the user, so that the user can identify and determine depression of the target object based on the eye movement classification results.

2. The method for identifying depression based on eye movement image recognition according to claim 1, characterized in that: The step of performing image grayscale processing on the eye monitoring image frame to obtain a corresponding eye grayscale image, and performing image filtering processing on the eye grayscale image to obtain a corresponding eye filtered grayscale image comprises: Determine the RGB three-color channel component value of each pixel in the eye monitoring image frame, and perform weighted summation of the RGB three-color channel component value of each pixel with a set weight ratio to obtain the grayscale value of each pixel, and obtain the eye grayscale image based on the grayscale value of each pixel; The eye grayscale image is processed by Gaussian filtering algorithm to obtain the corresponding eye filtered grayscale image.

3. The method for identifying depression based on eye movement image recognition according to claim 1, characterized in that: The step of performing image binarization processing on the eye filtered grayscale image to obtain a corresponding eye binary image includes: The Bernsen algorithm is used to perform image binarization on the eye filtered grayscale image to obtain the corresponding eye binary image.

4. The method for identifying depression based on eye movement image recognition according to claim 1, characterized in that: The method of performing closed edge detection on the eye binary image, determining the eye edge contour, and determining the eye corner points on both sides of the eye edge contour and the pupil image area within the eye edge contour includes: The Canny edge detection operator is used to perform edge detection on the eye binary image, and the contours of each closed edge are determined based on the edge detection results; A closed edge contour with an area within a set first area range and corner points on both sides is used as an eye edge contour, and the corner points on both sides of the eye edge contour are used as eye corner points on both sides; A closed edge contour whose area within the eye edge contour is within a set second area range is used as a pupil edge contour, and an image area within the pupil edge contour is used as a pupil image area.

5. The method for identifying depression based on eye movement image recognition according to claim 1, characterized in that: The step of constructing a first feature sequence based on the feature value sequence of each eye monitoring image frame includes: Taking the average of the feature values ​​in the same order in the feature value sequence of each eye monitoring image frame to obtain the feature average value of the corresponding order; Combine the average values ​​of each feature in order according to the corresponding ranking to obtain the first feature sequence {Z1, Z2, ..., Z n }, where Z n The average value of the features ranked by n.

6. The method for identifying depression based on eye movement image recognition according to claim 5, characterized in that: The step of constructing a second feature sequence based on the reference center coordinates of each eye monitoring image frame includes: The reference center coordinates of each eye monitoring image frame are combined according to the order of the corresponding eye monitoring image frame to obtain a second feature sequence {(X1, Y1), (X2, Y2), ..., (X m , Y m )}, where X m and Y m Respectively represent the reference center horizontal coordinate and the reference center vertical coordinate ranked as m.

7. The method for identifying depression based on eye movement image recognition according to claim 6, characterized in that: The constructing a feature matrix based on the first feature sequence and the second feature sequence includes: Using the first feature sequence {Z1, Z2, ..., Z n } and the second characteristic sequence {(X1, Y1), (X2, Y2), ..., (X m , Y m )}Construct feature matrix Among them, m=n.

8. The method for identifying depression based on eye movement image recognition according to claim 1, characterized in that: The eye movement classification model adopts a trained ResNet model.

9. A depression recognition system based on eye movement image recognition, characterized in that: It includes an image acquisition unit, a graying filter unit, a binary processing unit, an edge detection unit, a feature calibration unit, a coordinate determination unit, a feature arrangement unit, a matrix construction unit and a classification recognition unit, wherein: An image acquisition unit, used to acquire an eye monitoring video segment of a target object in a set test scene, and extract a plurality of sequentially arranged eye monitoring image frames from the eye monitoring video segment at intervals; A graying filter unit is used to perform image graying processing on the eye monitoring image frame to obtain a corresponding eye gray image, and perform image filtering processing on the eye gray image to obtain a corresponding eye filtered gray image; A binary processing unit, used for performing image binarization processing on the eye filtered grayscale image to obtain a corresponding eye binary image; An edge detection unit is used to perform closed edge detection on the eye binary image, determine the eye edge contour, and determine the eye corner points on both sides of the eye edge contour and the pupil image area within the eye edge contour; A feature calibration unit is used to connect the eye corner points on both sides to obtain a horizontal reference line, and select a number of reference points at equal intervals on the horizontal reference line, and construct a vertical reference line on each reference point, so that the vertical reference line is perpendicular to the horizontal reference line, and the two ends of the vertical reference line are respectively in contact with the eye edge contour; A coordinate determination unit, used to construct an image coordinate axis with one of the eye corner points as the origin and the horizontal reference line as the horizontal coordinate, and determine the coordinates of the center point of the pupil image area, and use the coordinates of the center point of the pupil image area as the reference center coordinates of the corresponding eye monitoring image frame; A feature arrangement unit, used to take the ratio of the longitudinal reference line to the transverse reference line of each reference point as its feature value, and arrange the feature values ​​of each reference point in sequence to obtain a feature value sequence of the corresponding eye monitoring image frame; a matrix construction unit, configured to construct a first feature sequence based on a feature value sequence of each eye monitoring image frame, construct a second feature sequence based on a reference center coordinate of each eye monitoring image frame, and construct a feature matrix based on the first feature sequence and the second feature sequence; The classification and recognition unit is used to input the feature matrix into a preset eye movement classification model for classification and recognition, obtain the corresponding eye movement classification results, and display the eye movement classification results to the user, so that the user can identify and judge the depression of the target object based on the eye movement classification results.

10. A depression recognition system based on eye movement image recognition, characterized in that: include: A memory for storing execution instructions; A processor is used to read the execution instruction stored in the memory, and execute the depression identification method based on eye movement image recognition according to any one of claims 1-8 according to the execution instruction.

Images