Eye movement data processing method and system for detection of autism spectrum disorder

Abstract

Disclosed in the present invention are an eye movement data processing method and system for detection of autism spectrum disorder. The method comprises: on the basis of different motion patterns, displaying a plurality of detection icons in a detection task, the moving detection icons guiding a subject to viewing a display module; then collecting an eye movement image sequence of eyes of the subject viewing the detection icons, extracting therefrom eye movement data, and grouping the eye movement data; calculating a gaze point displacement data set corresponding to each group of eye movement data, importing a motion function corresponding to each detection icon into a test coordinate system, and importing each gaze point displacement dataset into the test coordinate system; and determining which motion function each gaze point displacement dataset conforms to, so as to dynamically associate the gaze position of the subject with the detection icons, thereby capturing the pattern of the subject's gaze following the detection icons. The detection method of the present invention does not need to be used with high-precision and high-cost hardware, improving the general applicability to meet large-scale and fast detection screening scenarios.

Description

An eye movement data processing method and system for autism spectrum disorder detection TECHNICAL FIELD

[0001] The present application belongs to the technical field of image data processing, and in particular relates to an eye movement data processing method and system for autism spectrum disorder detection. BACKGROUND

[0002] Currently, the detection methods of autism spectrum disorder are divided into two categories: the first category detects the three core features of autism spectrum disorder and the behavior patterns specific to autism spectrum disorder based on the core features; the second category detects the physiological indicators (including but not limited to eye movement parameters, brain structure and connectivity related parameters, etc.) that are different between autism disorder patients and normal people.

[0003] The existing various questionnaires and scales (such as "Erxin Scale-II" and "CARS") for screening and detecting autism spectrum disorder are based on the first category. They design scene questionnaires for detection, and cooperate with actual interaction and physical operation to achieve the purpose of assessing the disease condition. On this basis, some existing devices reproduce the required detection scene in the form of visual scene images and videos, and optimize and simplify the manual operation with an intelligent method. They cooperate with eye trackers to capture the corresponding gaze positions and submit the detection results to doctors as a basis for assisting doctors in diagnosis. However, the detection form of traditional scales needs to be completed for a very long time, and doctors need to maintain professional interaction and observation with patients and their relatives throughout the process, which is high in work intensity. It is not conducive to carrying out large-scale and rapid screening work, and the efficiency of diagnosis work is also relatively low. The auxiliary diagnosis device cooperating with the eye tracker needs to perform an eye calibration process of about 15 minutes before each detection of each patient, and then the detection process can begin. Since the exact gaze position is needed to determine the actual viewing content of the patient, the accuracy of the eye tracker is also very high. The above two reasons also make the first category not conducive to carrying out large-scale and rapid screening work. At the same time, the calibration process in the early stage also requires high cooperation of the patient, which increases the complexity of the doctor's operation, making it more difficult to complete the detection. Low-month-old infants are difficult to cooperate with the calibration, which easily makes the low-month-old infants lose the opportunity for early intervention. TECHNICAL PROBLEM

[0004] The autism spectrum disorder patients have differences in the eye jump parameters (such as eye jump amplitude, eye jump latency / eye jump reaction time, eye jump peak speed), eye movement parameters (such as mean, standard deviation / standard error, effect value), pupil size and other values when looking at specific objects, scenes, actions and the like. On the basis of the auxiliary diagnosis device described in the foregoing, higher-precision eye movement instruments are used to acquire and analyze relevant physiological information and give statistical data, and the accuracy and reliability of the detection results are improved by comprehensively judging the reaction of the patient to the detection scene. The shortcomings of the second type are the same as those of the first type. At the same time, higher precision represents higher cost, which further magnifies the above shortcomings. Technical solutions

[0005] The purpose of the present application is to provide a more efficient eye movement data processing method and system for autism spectrum disorder detection. Based on the three core characteristics of autism spectrum disorder (social impairment, communication impairment, narrow interest and behavioral stereotypy) and the specific behavior patterns (joint attention paradigm, name calling reaction paradigm, non-social sound stimulus behavior paradigm and the like), a detection task is designed, a display device is used to display each detection icon in the detection task with different motion rules, eye movement image sequences of the testee watching the detection icons are collected, the eye movement image sequences are analyzed, the motion rules of the detection icons are compared, the following of the testee's eye movement to the detection icons is judged, and the test results are obtained to assist doctors in diagnosis.

[0006] To solve the above technical problems, the following technical solutions are adopted in the present application:

[0007] An eye movement data processing method for autism spectrum disorder detection is proposed, comprising:

[0008] S1, displaying a group of moving detection icons on a display module; wherein the group of detection icons contains at least two detection icons, and each detection icon moves according to a different motion function;

[0009] S2, collecting eye movement image sequences of the testee watching the display module, and analyzing to obtain eye movement position data, constructing an eye movement data scatter plot with time as the horizontal axis and eye movement position as the vertical axis;

[0010] S3, grouping the eye movement data based on the eye movement data scatter plot;

[0011] S4, calculating the line-of-sight landing point displacement data set of each data group, and constructing a line-of-sight landing point displacement data scatter plot;

[0012] S5, constructing a test coordinate system, and importing the motion function corresponding to the detection icon into the test coordinate system;

[0013] S6, sequentially import each group of sight line landing point displacement data set into the test coordinate system and compare with each motion function graph to find the motion function that each sight line landing point displacement data set conforms to;

[0014] S7, classify each group of sight line landing point displacement data set and output statistical results.

[0015] In some embodiments of the present application, step S3 specifically comprises:

[0016] A constant k is set, and two adjacent data in the horizontal axis direction are sequentially taken from the origin of the eye movement data scatter plot coordinate system; the constant k is a critical value between the eye movement amplitude and the saccade amplitude when tracking the gaze, and is reflected in the displacement amount in the collected eye movement image;

[0017] The vertical coordinate values of the two data are obtained, and the absolute value Δk of the difference between them is calculated;

[0018] Δk is compared with the constant k, if Δk < k, the two data are classified into the same group of data, if Δk > k, a is taken as the segmentation point of the data set, a is classified into the previous group of data, and b is classified into a new group of data.

[0019] In some embodiments of the present application, the calculation of the sight line landing point displacement data set of each data group in step S4 specifically comprises:

[0020] The position coordinates of the eye movement data are multiplied by the constant to obtain the sight line landing point displacement data set; wherein, , is the displacement amount of a single detection icon, is the corresponding iris displacement amount in the eye movement image sequence.

[0021] In some embodiments of the present application, step S5 specifically comprises:

[0022] The center of the detection icon is taken as the origin of the test coordinate system, the motion function is the motion trajectory of the origin, the distance between the left edge frame of the detection icon and the origin and the distance between the right edge frame and the origin are set as a, the motion function of the left edge of the detection icon is obtained and the motion function of the right edge frame , is the number of the motion function;

[0023] The motion function of the left edge and the motion function of the right edge are imported into the test coordinate system, and the range surrounded by the two motion functions constitutes the actual motion area of the entire frame of the detection icon.

[0024] In some embodiments of the present application, step S6 specifically comprises:

[0025] sequentially import each set of line-of-sight landing point displacement data into the test coordinate system;

[0026] For each set of line-of-sight landing point displacement data, compare with the motion region corresponding to each detection icon to find the number of data points of each set of line-of-sight landing point displacement data falling into each motion region .

[0027] The maximum value corresponds to the motion function that the line-of-sight landing point displacement data set conforms to.

[0028] An eye movement data processing system for detecting autism spectrum disorder is proposed, comprising:

[0029] A display module for displaying a set of moving detection icons; wherein the set of detection icons contains at least two detection icons, and each detection icon moves according to a different motion function;

[0030] A collection module for collecting eye movement image sequences of a test subject when watching the display module;

[0031] An operation module for processing eye movement data according to the following steps:

[0032] Parse the eye movement position data to construct an eye movement data scatter plot with time as the horizontal axis and eye movement position as the vertical axis;

[0033] Group the eye movement data based on the eye movement data scatter plot;

[0034] Calculate the line-of-sight landing point displacement data set of each data group to construct a line-of-sight landing point displacement data scatter plot;

[0035] Construct a test coordinate system and import the motion function corresponding to the detection icon into the test coordinate system;

[0036] Import each set of line-of-sight landing point displacement data into the test coordinate system and compare with each motion function graph to find the motion function that each set of line-of-sight landing point displacement data conforms to;

[0037] Classify each set of line-of-sight landing point displacement data and output the statistical results.

[0038] In some embodiments of the present application, when the operation module groups the eye movement data based on the eye movement data scatter plot, it includes:

[0039] Set a constant k, starting from the origin of the eye movement data scatter plot coordinate system, take two adjacent data in the horizontal axis direction in turn; the constant k is: the critical value between the eye movement amplitude and the saccade amplitude when tracking the gaze, which is reflected in the displacement amount in the collected eye movement image; ​

[0040] Obtaining longitudinal coordinate values of two data, calculating absolute value of difference between the two data Δk;

[0041] Comparing Δk with constant k, if Δk < k, then the two data are classified into the same data group, if Δk > k, then a is classified into the previous data group and b is classified into a new data group.

[0042] In some embodiments of the present application, the operation module specifically comprises the following steps when calculating the gaze landing point displacement data set of each data group:

[0043] Multiplying the position coordinates of the eye movement data with constant to obtain the gaze landing point displacement data set; wherein, , is the displacement of a single detection icon, is the corresponding iris displacement in the eye movement image sequence.

[0044] In some embodiments of the present application, the operation module imports the motion function corresponding to the detection icon into the test coordinate system, specifically comprising:

[0045] Taking the center of the detection icon as the origin of the test coordinate system, the motion trajectory of the motion function as the origin, and setting the distance between the left edge frame of the detection icon and the origin and the distance between the right edge frame and the origin as a, the motion function of the left edge of the detection icon is obtained and the motion function of the right edge frame , is the number of the motion function;

[0046] Importing the motion function of the left edge and the motion function of the right edge into the test coordinate system, the range surrounded between the two motion functions constitutes the actual motion area of the entire frame of the detection icon.

[0047] In some embodiments of the present application, the operation module finds the motion function that each gaze landing point displacement data set conforms to, specifically comprising:

[0048] Importing each gaze landing point displacement data set into the test coordinate system in turn;

[0049] Comparing each gaze landing point displacement data set with each motion area corresponding to each detection icon respectively, finding the number of data points of each gaze landing point displacement data set falling into each motion area ;

[0050] Taking the motion function corresponding to the maximum value as the motion function that the gaze landing point displacement data set conforms to. Beneficial effects ​

[0051] Compared with the prior art, the advantages and positive effects of this application are as follows: The eye movement data processing method and system for autism spectrum disorder detection proposed in this application, based on the core characteristics and unique behavioral patterns of autism spectrum disorder, designs a test task based on observation and detection icons to mine the eye movement patterns of the test subjects, thereby enabling the test subjects to determine whether they have autism spectrum disorder based on the mined eye movement patterns. In the test, a display module was used to display multiple detection icons in the detection task with different motion patterns. Based on the motion of the detection icons, the test subject was guided to look at the display module. Then, eye-tracking image sequences of the test subject's eyes looking at the detection icons were collected. Eye-tracking data was extracted from the eye-tracking image sequences and grouped. The displacement dataset of the gaze point corresponding to each group of eye-tracking data was calculated. The motion function corresponding to each detection icon was imported into the test coordinate system, and each displacement dataset of the gaze point was imported into the test coordinate system. It was determined which motion function each displacement dataset of the gaze point conformed to, thereby dynamically associating the test subject's gaze position with the detection icon. This achieved the capture of the pattern of the test subject's gaze following the detection icon. Compared with the existing detection methods that use eye trackers, the method of this invention can determine the test subject's gaze position in real time, eliminate the instrument calibration process, and achieve the effect of no adjustment by the test subject throughout the process. It can effectively improve detection efficiency and reduce the difficulty of detection operation. At the same time, the detection method of this invention does not require the use of high-precision, high-cost hardware, improves versatility, and can meet the screening scenarios of large-scale and rapid detection.

[0052] Other features and advantages of this application will become clearer after reading the detailed description of the embodiments in conjunction with the accompanying drawings. Attached Figure Description

[0053] Figure 1 is a schematic diagram of the steps of the eye movement data processing method for autism spectrum disorder detection proposed in this invention;

[0054] Figure 2 shows an example of the detection icon group displayed with different motion functions in this invention;

[0055] Figure 3 is an example of an eye-tracking data scatter plot constructed in this invention;

[0056] Figure 4 is an example of the grouped scatter plot of eye movement data shown in Figure 3 in this invention;

[0057] Figure 5 illustrates the relationship between the amount of iris movement of the human eye and the displacement of the detection icon being viewed by the eye.

[0058] Figure 6 is an example of a scatter plot of eye movement data displacement data shown in Figure 4 of this invention.

[0059] Figure 7 is a schematic diagram of the motion function corresponding to a detection icon in the present invention after being imported into the test coordinate system;

[0060] Figure 8 shows an example of a detection icon in this invention;

[0061] Figure 9 is an example of the motion area of ​​the detection icon shown in Figure 8 in the test coordinate system of the present invention;

[0062] Figure 10 is a schematic diagram of importing a line-of-sight point displacement dataset into the test coordinate system in this invention;

[0063] Figure 11 is a schematic diagram comparing the motion region of a detection image with the line-of-sight displacement dataset shown in Figure 10.

[0064] Figure 12 is a schematic diagram comparing the motion region of the line-of-sight displacement dataset shown in Figure 10 and another detection image. The best embodiment of the present invention

[0065] The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0066] This invention aims to propose an eye-tracking data processing method for auxiliary diagnosis of autism spectrum disorder. The entire system includes a display module, an acquisition module, an operation module, a calculation module, and an output module. The display module displays the detection icons, and the content of the display is controlled by the calculation module. The acquisition module acquires eye-tracking image sequences of the test subject while viewing the detection icons. The operation module controls the start, stop, and pause operations of the detection. The calculation module drives each module, runs the algorithm, and obtains the data processing results. The output module outputs the data processing results electronically or in paper format.

[0067] Referring to Figure 1, the eye movement data processing method for autism spectrum disorder detection proposed in this invention includes the following steps:

[0068] S1: Display a group of motion detection icons on the display module; the group of detection icons contains at least two detection icons, each of which moves according to a different motion function.

[0069] For example, multiple detection icons in a detection icon group are displayed horizontally back and forth on the display module with different motion functions. For example, as shown in Figure 2, the detection icon for the human head image moves back and forth from left to right at a first set rate, while the detection icon for the train head image moves back and forth from right to left at a second set rate.

[0070] In this process, multiple detection images are labeled as detection icon 1, detection icon 2, ..., detection icon n; the corresponding motion functions are denoted as... , ... .

[0071] S2: Collect the eye movement image sequence of the tested person when viewing the display module, parse to obtain the eye movement position data, and construct a scatter plot of eye movement data with time as the horizontal axis and eye movement position as the vertical axis.

[0072] Collect the eye movement image sequence of the tested person during the viewing of the detection icon on the display module through the acquisition module, parse the eye movement image sequence to obtain the position data (horizontal and / or vertical) of the eye movement, establish a coordinate system for the scatter plot of eye movement data with time as the horizontal axis and position data as the vertical axis, import the position data of the eye movement into the coordinate system according to the acquisition time of the image sequence, and obtain the scatter plot of eye movement data. As shown in Figure 3, it is the scatter plot of eye movement data obtained by importing the horizontal position data of the eye movement into the coordinate system according to the acquisition time.

[0073] S3: Group the eye movement data based on the scatter plot of eye movement data.

[0074] Set a constant k. Starting from the origin of the coordinate system of the scatter plot of eye movement data, sequentially take two adjacent data a and b in the horizontal axis direction (that is, the eye movement data corresponding to two adjacent time points), obtain the vertical coordinate values of the two data, calculate the absolute value Δk of the difference between the two, compare Δk with the constant k. If Δk < k, then classify the two data into the same group of data. If Δk > k, then take a as the segmentation point of the data set, classify a into the previous group of data, and classify b into a new group of data.

[0075] The purpose of this step is to optimize the fitting degree of the fitting curve in subsequent operations and reduce interference.

[0076] The actual meaning of the constant k is: the critical value between the eye movement amplitude and saccade amplitude during tracking gaze, manifested as the displacement amount in the collected eye movement images.

[0077] The scatter plot after grouping is shown in Figure 4. Assume that N groups of eye movement data are finally separated.

[0078] S4: Calculate the data set of the line-of-sight landing point displacement for each data group and construct a scatter plot of the line-of-sight landing point displacement data.

[0079] As shown in Figure 5, within a unit time, there is a proportional relationship between the movement amount L1 of the iris of the human eye itself and the displacement amount L2 of the detection icon being gazed at by the eye. In the figure, o is the position of the eye axis, L3 is the distance between the iris and the eye axis, and L4 is the distance between the iris and the display module. There is . Similarly, there is also a proportional relationship between the iris displacement amount L0 in the eye movement image sequence and the movement amount L1 of the iris of the human eye itself. Therefore, there is a proportional relationship among the detection icon displacement amount L2, the iris displacement amount L0 in the eye movement image sequence, and the movement amount L1 of the iris of the human eye itself. By measuring the displacement amount L2 of a single detection icon and the corresponding iris displacement amount L0 in the eye movement image sequence, the ratio between the two can be obtained. .

[0080] Based on the above principles, the eye-tracking dataset Multiplying the position coordinates by the constant J yields the displacement dataset of the line-of-sight point. or Based on the new dataset, a scatter plot was reconstructed to obtain a scatter plot of the line-of-sight displacement data, as shown in Figure 6.

[0081] S5: Construct a test coordinate system and import the motion function corresponding to the detection icon into the test coordinate system.

[0082] Construct a test coordinate system and plot the motion function of each detection icon (1, 2, ..., n). , ... Importing the constructed test coordinate system, as shown in an example in Figure 7.

[0083] S6: Import each set of line-of-sight displacement datasets into the test coordinate system and compare them with the graphs of each motion function to find the motion function that each line-of-sight displacement dataset conforms to.

[0084] Taking the division of eye-tracking data into four groups in step S3 as an example, respectively using... , , , It means, first Import the data set into the test coordinate system, and determine which range of the motion function graph each data point falls within. Assume... If the data set contains 10 data points, determine which region of the motion function graph each of the 10 data points falls within, and count the number of data points falling within the same region of the motion function graph. Set a threshold m; when the number of data points falling within the same region of the motion function graph exceeds the threshold m, a decision is made. The data points in the data set conform to this motion function, corresponding to the test subject's gaze following the movement of the detection icon that follows this motion function. Then... Import the data from the data set into the test coordinate system and repeat the above steps to determine which motion function it conforms to; then... Data group Import the data set into the test coordinate system and repeat the above steps to determine which motion function each data set conforms to.

[0085] S7: Classify the datasets of line-of-sight displacements for each group and output the statistical results.

[0086] In other words, eye-tracking data from people viewing the same detection icon are collected into a set and the results are output.

[0087] Step S6 has already determined which motion function each line-of-sight point displacement dataset conforms to, for example... The data in the data set conforms to the motion function. , The data in the data set conforms to the motion function. , The data in the data set conforms to the motion function. , The data in the data set conforms to the motion function. Then Data sets and motion functions They are categorized and their binding relationships are output. , and Data sets and motion functions They are categorized and their binding relationships are output.

[0088] Based on the core characteristics and unique behavioral patterns of autism spectrum disorder, the present invention designs a test task based on observation and detection icons to mine the eye movement patterns of test subjects, thereby enabling test subjects to determine whether they have autism spectrum disorder based on the mined eye movement patterns. In the test, a display module was used to display multiple detection icons in the detection task with different motion patterns. Based on the motion of the detection icons, the test subject was guided to look at the display module (to test the test subject's interest in and ability to follow each detection icon). Then, eye-tracking image sequences of the test subject's eyes looking at the detection icons were collected. Eye-tracking data was extracted from the eye-tracking image sequences and grouped. The displacement dataset of the gaze point corresponding to each group of eye-tracking data was calculated. The motion function corresponding to each detection icon was imported into the test coordinate system, and each displacement dataset of the gaze point was imported into the test coordinate system. It was determined which motion function each displacement dataset of the gaze point conformed to, thereby dynamically associating the test subject's gaze position with the detection icon. This achieved the capture of the pattern of the test subject's gaze following the detection icon. Compared with the existing detection methods that use eye trackers, the method of this invention can determine the test subject's gaze position in real time, eliminate the instrument calibration process, and achieve the effect of no adjustment by the test subject throughout the process. It can effectively improve detection efficiency and reduce the difficulty of detection operation. At the same time, the detection method of this invention does not require the use of high-precision, high-cost hardware, improves versatility, and can meet the screening scenarios of large-scale and rapid detection.

[0089] The steps S5 and S6 of the eye-tracking data processing method proposed in this invention will be described in detail below with a specific embodiment.

[0090] Taking the detection icon shown in Figure 8 as an example, the center of the detection icon is set as the origin of the test coordinate system. Let be the trajectory of the movement from the origin. Let 'a' be the distance between the left edge of the detection icon and the origin, and 'a' be the distance between the right edge and the origin. Then the motion function for the left edge of the detection icon is: The motion function of the right edge of the image is Import these two functions into the test coordinate system. The area enclosed between the two functions is the actual motion area of ​​the entire screen of the detection icon, as shown in Figure 9.

[0091] Each segment of the line-of-sight displacement dataset is sequentially imported into the test coordinate system and compared with the motion area corresponding to each detection icon. As shown in Figure 10, the first segment of the line-of-sight displacement dataset is imported into the test coordinate system and compared with the motion area of ​​detection icon 1. In each dataset, the data points are represented by multiple tables. Keeping the horizontal axis (time axis) values ​​unchanged, append a constant value to the vertical axis data for all data points. ,get This achieves the overall translation of the data points in the dataset along the vertical axis in the test coordinate system; whereby... The value satisfies the following condition: as many data points as possible in the current dataset are included in the motion region of the current detection icon.

[0092] After the above conditions are met, the total number of data points in the statistical data set is denoted as . , Let be the index of the data set of the line-of-sight displacement; then count the number of data points contained in the j-th motion region, denoted as . .

[0093] In the embodiment shown in Figure 11, the first dataset contains a total of 11 data points. After uniformly shifting all points in the dataset vertically, a maximum of 9 data points can be included in the motion region of detection icon 1. Therefore, the first dataset... , Similarly, as shown in Figure 12, by comparing the motion region of the first dataset segment with that of detection icon 2, we can obtain... , .

[0094] Repeat the above steps; after comparing each dataset segment, multiple results can be obtained. Value and a value.

[0095] Pick Maximum value ,like Then this dataset segment and the detection icon The overlap of the movement areas is highest at this point, at which time the test subject's gaze is judged to follow the detection icon. It did not follow other detection icons.

[0096] In some embodiments of the present invention, a constant is set. Used for Perform a judgment if Then, it is determined that the test subject's gaze is following the detection icon within the current time period of the dataset. ,like If the current dataset is within a certain time period, it is determined that the test subject's gaze did not follow any detection icon.

[0097] In some embodiments of the present invention, the statistical data output in step S7 can be visualized in the form of charts or other forms, so that testers can intuitively understand the classification results.

[0098] Based on the eye-tracking data processing method for autism spectrum disorder detection proposed above, this invention also proposes an eye-tracking data processing system for autism spectrum disorder detection, comprising:

[0099] The display module is used to display a group of motion detection icons; wherein the group of detection icons contains at least two detection icons, and each detection icon moves according to a different motion function.

[0100] The acquisition module is used to acquire eye-tracking image sequences when the test subject views the display module.

[0101] The computation module processes eye-tracking data in the following steps: parsing the eye-tracking position data and constructing a scatter plot of the eye-tracking data with time as the horizontal axis and eye-tracking position as the vertical axis; grouping the eye-tracking data based on the scatter plot; calculating the gaze-point displacement dataset for each data group and constructing a gaze-point displacement dataset scatter plot; constructing a test coordinate system and importing the motion function corresponding to the detection icon into the test coordinate system; sequentially importing each group of gaze-point displacement datasets into the test coordinate system and comparing them with the graphs of each motion function to find the motion function that each gaze-point displacement dataset conforms to; classifying each group of gaze-point displacement datasets and outputting the statistical results.

[0102] When the operation module groups eye movement data based on the scatter plot of eye movement data, it includes: setting a constant k, starting from the origin of the coordinate system of the eye movement data scatter plot, and successively taking two adjacent data in the horizontal axis direction; the constant k is: the critical value between the eye movement amplitude and the saccade amplitude during tracking gaze, which is the displacement amount shown in the collected eye movement image; obtaining the vertical coordinate values of the two data, and calculating the absolute value Δk of the difference between the two; comparing Δk with the constant k, if Δk < k, then classify the two data into the same group of data, if Δk > k, then take a as the segmentation point of the data set, classify a into the previous data group, and classify b into a new data group.

[0103] When the operation module calculates the line-of-sight landing point displacement data set of each data group, it specifically includes: multiplying the position coordinates of the eye movement data by the constant to obtain the line-of-sight landing point displacement data set; where , is the displacement amount of a single detection icon, is the iris displacement amount corresponding to the eye movement image sequence.

[0104] When the operation module imports the motion function corresponding to the detection icon into the test coordinate system, it specifically includes: taking the center of the detection icon as the origin of the test coordinate system, and the motion function as the motion trajectory of the origin, setting the distance between the left edge picture of the detection icon and the origin and the distance between the right edge picture and the origin as a, and obtaining the motion function of the left edge of the detection icon and the motion function of the right edge picture, is the number of the motion function; importing the motion function of the left edge and the motion function of the right edge into the test coordinate system, and the range enclosed by the two motion functions constitutes the actual motion area of the entire picture of the detection icon.

[0105] When the operation module finds the motion function that each line-of-sight landing point displacement data set conforms to, it specifically includes: successively importing each group of line-of-sight landing point displacement data sets into the test coordinate system; comparing each group of line-of-sight landing point displacement data sets with the motion area corresponding to each detection icon respectively, and finding the number of data points where each group of line-of-sight landing point displacement data sets fall into each motion area; judging the motion function corresponding to the maximum value as the operating function that the line-of-sight landing point displacement data set conforms to.

[0106] The present invention also proposes an autism spectrum disorder detection device, which is configured with the above-mentioned eye movement data processing system, runs the above-mentioned eye movement data processing method, outputs the eye movement pattern data of the tested person, and the tester can combine the core characteristics and behavior patterns of autism spectrum disorder, and assist in judging whether the tested person has autism spectrum disorder according to the eye movement pattern data of the tested person.

[0107] It should be noted that, in the specific implementation process, the above-mentioned control part can be implemented by a hardware processor executing computer-executable instructions in software form stored in memory, which will not be elaborated here. The programs corresponding to the actions performed by the above control circuit can all be stored in the computer-readable storage medium of the system in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0108] The computer-readable storage media mentioned above may include volatile memory, such as random access memory; may also include non-volatile memory, such as read-only memory, flash memory, hard disk or solid-state drive; and may also include combinations of the above types of memory.

[0109] The term "processor" as mentioned above can also refer to a collective of multiple processing elements. For example, a processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor, and it can also be a special-purpose processor.

[0110] It should be noted that the above description is not intended to limit the present invention, and the present invention is not limited to the examples given above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A method for processing eye-tracking data for autism spectrum disorder detection, characterized in that, Including: S1, displaying a group of moving detection icons on the display module; among which, the detection icon group includes at least two detection icons, and each detection icon moves according to a different motion function; S2, collecting the eye movement image sequence of the tested person when watching the display module, parsing to obtain the eye movement position data, taking time as the horizontal axis and the eye movement position as the vertical axis to construct an eye movement data scatter plot; S3, grouping the eye movement data based on the eye movement data scatter plot; S4, calculating the line-of-sight landing point displacement data set of each data group and constructing a line-of-sight landing point displacement data scatter plot; S5, constructing a test coordinate system and importing the motion function corresponding to the detection icon into the test coordinate system; S6, sequentially importing each group of line-of-sight landing point displacement data sets into the test coordinate system to compare with each motion function graph, and finding the motion function that each line-of-sight landing point displacement data set conforms to; S7, classifying each group of line-of-sight landing point displacement data sets and outputting the statistical result.

2. The eye-tracking data processing method for autism spectrum disorder detection according to claim 1, characterized in that, Step S3 specifically includes: Setting a constant k, starting from the origin of the eye movement data scatter plot coordinate system, sequentially taking two adjacent data in the horizontal axis direction; the constant k is: the critical value between the eye movement amplitude and the saccade amplitude during tracking gaze, manifested as the displacement amount in the collected eye movement image; Obtaining the longitudinal coordinate values of the two data and calculating the absolute value Δk of the difference between the two; Comparing Δk with the constant k. If Δk < k, then classify the two data into the same group of data. If Δk > k, then taking a as the segmentation point of the data set, classifying a into the previous group of data and b into a new group of data.

3. The eye-tracking data processing method for autism spectrum disorder detection according to claim 1, characterized in that, In step S4, calculating the line-of-sight landing point displacement data set of each data group specifically includes: The position coordinates of the eye-tracking data and constants Multiplying these yields the dataset of line-of-sight displacements; where, ， This represents the displacement of a single detection icon. It is the iris displacement amount corresponding to the eye movement image sequence.

4. The eye-tracking data processing method for autism spectrum disorder detection according to claim 1, characterized in that, Step S5 of importing the motion function corresponding to the detection icon into the test coordinate system specifically includes: Using the center of the detection icon as the origin of the test coordinate system and the motion function as the motion trajectory of the origin, let 'a' be the distance between the left edge of the detection icon and the origin, and the distance between the right edge of the detection icon and the origin. Then, the motion function of the left edge of the detection icon is obtained. Motion function of the right edge of the image ， It is the number of the motion function; Importing the motion function of the left edge and the motion function of the right edge into the test coordinate system, and the range enclosed between the two motion functions constitutes the actual motion area of the entire picture of the detection icon.

5. The eye-tracking data processing method for autism spectrum disorder detection according to claim 4, characterized in that, Step S6 specifically includes: Sequentially importing each group of line-of-sight landing point displacement data sets into the test coordinate system; For each set of gaze point displacement datasets, compare it with the motion region corresponding to each detection icon to find the number of data points in each set of gaze point displacement datasets that fall into each motion region. ； Will Judging the motion function corresponding to the maximum value as the operating function that the line-of-sight landing point displacement data set conforms to.

6. An eye-tracking data processing system for detecting autism spectrum disorder, characterized in that, Including: A display module for displaying a group of moving detection icons; among which, the detection icon group includes at least two detection icons, and each detection icon moves according to a different motion function; A collection module for collecting the eye movement image sequence of the tested person when watching the display module; An operation module for processing eye movement data according to the following steps: Parsing to obtain the eye movement position data, taking time as the horizontal axis and the eye movement position as the vertical axis to construct an eye movement data scatter plot; Grouping the eye movement data based on the eye movement data scatter plot; Calculating the line-of-sight landing point displacement data set of each data group and constructing a line-of-sight landing point displacement data scatter plot; Constructing a test coordinate system and importing the motion function corresponding to the detection icon into the test coordinate system; Sequentially importing each group of line-of-sight landing point displacement data sets into the test coordinate system to compare with each motion function graph, and finding the motion function that each line-of-sight landing point displacement data set conforms to; Classifying each group of line-of-sight landing point displacement data sets and outputting the statistical result.

7. The eye-tracking data processing system for autism spectrum disorder detection according to claim 6, characterized in that, When grouping eye movement data based on the scatter plot of eye movement data, the operation module includes: Setting a constant k, starting from the origin of the coordinate system of the eye movement data scatter plot, and successively taking two adjacent data in the horizontal axis direction; the constant k is: the critical value between the eye movement amplitude and the saccade amplitude during tracking and fixation, manifested as the displacement amount in the collected eye movement image; Obtaining the vertical coordinate values of the two data, and calculating the absolute value Δk of the difference between the two; Comparing Δk with the constant k. If Δk < k, the two data are classified into the same group of data. If Δk > k, a is used as the segmentation point of the data set, a is classified into the previous group of data, and b is classified into a new group of data.

8. The eye-tracking data processing system for autism spectrum disorder detection according to claim 6, characterized in that, When calculating the line-of-sight landing point displacement data set of each data group, the operation module specifically includes: The position coordinates of the eye-tracking data and constants Multiplying these yields the dataset of line-of-sight displacements; where, ， This represents the displacement of a single detection icon. Is the iris displacement amount corresponding to the eye movement image sequence.

9. The eye-tracking data processing system for autism spectrum disorder detection according to claim 6, characterized in that, The operation module imports the motion function corresponding to the detection icon into the test coordinate system, specifically including: Using the center of the detection icon as the origin of the test coordinate system and the motion function as the motion trajectory of the origin, let 'a' be the distance between the left edge of the detection icon and the origin, and the distance between the right edge of the detection icon and the origin. Then, the motion function of the left edge of the detection icon is obtained. Motion function of the right edge of the image ， Is the number of the motion function; Importing the motion functions of the left edge and the right edge into the test coordinate system, and the range enclosed between the two motion functions constitutes the actual motion area of the entire picture of the detection icon.

10. The eye-tracking data processing system for autism spectrum disorder detection according to claim 9, characterized in that, When the operation module finds the motion function that each line-of-sight landing point displacement data set conforms to, it specifically includes: Successively importing each group of line-of-sight landing point displacement data sets into the test coordinate system; For each set of gaze point displacement datasets, compare it with the motion region corresponding to each detection icon to find the number of data points in each set of gaze point displacement datasets that fall into each motion region. ； Will The motion function corresponding to the maximum value is judged as the operation function that the line-of-sight landing point displacement data set conforms to.

Images