Attention detection system and method based on pupil size fluctuation characteristics

Abstract

The invention discloses an attention detection system and method based on pupil size fluctuation characteristics, and relates to the technical field of attention detection.The key points of the technical scheme are that pupil diameter data of a detected person in a resting state are collected through an eye tracker; processing the pupil diameter data, deleting blink data points, and supplementing data missing values by adopting linear interpolation to obtain interpolated pupil diameter data; converting the pupil diameter data after interpolation, converting a time domain signal into a frequency domain signal through a Fourier transform algorithm, and calculating to obtain a power spectrum; performing aperiodic index fitting on the power spectrum to generate a spectrum slope; and evaluating the attention level of the detected person according to the spectrum slope. High-precision pupil data are collected through the non-invasive glasses type eye tracker, convenient, objective and high-sensitivity attention assessment is achieved in combination with spectrum slope analysis, and an innovative solution is provided for clinical diagnosis, educational intervention and occupational health management.

Description

Technical Field

[0001] The present invention relates to the field of attention detection technology, and more specifically, to an attention detection system and method based on pupil size fluctuation characteristics. Background Art

[0002] In today's society, attention problems are a widespread concern. According to statistics, 5.29% to 7.2% of children suffer from attention deficit hyperactivity disorder (ADHD), and approximately half of these individuals continue to struggle with attention issues as adults. Attention problems not only affect the learning, behavior, and conduct of children and adolescents, but can also lead to poor work performance and low incomes in adults, as well as personality and emotional control issues. In specialized professions requiring prolonged concentration, such as drivers, security inspectors, and pilots, inattention can lead to serious personal and financial losses. Therefore, conducting scientific and accurate attention testing and assessments is of great significance.

[0003] Currently, there are many methods for existing attention detection technologies, but they still have the following shortcomings:

[0004] Contact physiological monitoring technology: It mainly relies on brain waves. The test subject needs to wear equipment such as an electrode cap, which may cause scalp discomfort and allergies to conductive media. In addition, EEG detection is easily affected by interference signals such as motion artifacts. It requires a high degree of cooperation from the test subject and requires professional personnel to install electrodes and interpret signals. The process is cumbersome and time-consuming.

[0005] Behavioral analysis: This approach assesses attention by recording the subject's overt behavior while completing specific tasks, such as mouse tracking. However, this method only captures overt behavior and cannot detect implicit cognitive states. Furthermore, it is limited by device type and tasks and cannot be applied to non-operational scenarios. Furthermore, it requires pre-training a personalized behavioral model for each subject, resulting in poor generalization.

[0006] Traditional psychological scales and interview assessments are highly subjective, easily influenced by deceptive responses from subjects, have limited scope of application, and are difficult to quantify and accurately measure attention. Furthermore, different diagnosticians, parents, and teachers use different criteria, leading to large deviations in diagnostic results and a lack of correlation with objective physiological indicators, resulting in insufficient diagnostic accuracy.

[0007] Recent studies have shown that pupil size is an important physiological indicator of the brain's wakefulness and is closely related to the level of attention. The dynamic changes in pupil diameter over time can reflect fluctuations in an individual's attention state. A large number of studies have shown that patients with attention deficit hyperactivity disorder (ADHD) have developmental and functional abnormalities in the nervous system involved in attention regulation and eye muscle control, suggesting that changes related to eye muscles may be one of the important characteristics of this type of disease. At the same time, the measurement of pupil size fluctuation involves periodic and non-periodic components. The non-periodic component is manifested as the characteristic that the power spectrum energy decreases with increasing frequency. In the resting state, the power spectrum slope of the non-periodic component of the neural signal is related to the excitation-inhibition balance of the nervous system and is also related to the level of attention. In other words, the non-periodic component of pupil size fluctuation can be used as an indicator of attention. However, existing technologies have not yet effectively utilized this physiological feature to achieve attention detection.

[0008] Therefore, the present invention aims to provide an attention detection system and method based on pupil size fluctuation characteristics to solve the problems existing in existing attention detection technology. Summary of the Invention

[0009] The purpose of the present invention is to provide an attention detection system and method based on pupil size fluctuation characteristics. The present invention collects high-precision pupil data through a non-invasive glasses-type eye tracker, combined with spectral slope analysis, to achieve convenient, objective and highly sensitive attention assessment, providing innovative solutions for clinical diagnosis, educational intervention and occupational health management.

[0010] The above technical objectives of the present invention are achieved through the following technical solutions: an attention detection system and method based on pupil size fluctuation characteristics, including a data acquisition module, a data processing module, a spectrum analysis module and a result evaluation module;

[0011] The data acquisition module is used to capture the eye image of the subject in real time, calculate the pupil diameter data and transmit it to the data processing module;

[0012] The data processing module is used to receive and process pupil diameter data transmitted by the data acquisition module, delete blinking data points, supplement missing data values ​​using linear interpolation, and then use Fourier transform algorithm to convert time domain signals into frequency domain signals and transmit them to the spectrum analysis module;

[0013] The spectrum analysis module is used to receive the frequency domain signal transmitted by the data processing module and calculate the power spectrum, and then generate the spectrum slope through non-periodic exponential fitting and transmit it to the result evaluation module;

[0014] The result evaluation module is used to receive the spectrum slope transmitted by the spectrum analysis module to evaluate the attention level of the person being measured.

[0015] The present invention is further configured as follows: the data processing module includes a data pre-processing unit and a signal conversion unit;

[0016] The data preprocessing unit is used to preprocess the pupil diameter data by deleting the data points of blinking and then supplementing the missing data values ​​by linear interpolation to obtain the pupil diameter data after interpolation;

[0017] The signal conversion unit is used to convert the interpolated pupil diameter data and convert the time domain signal into a frequency domain signal through a Fourier transform algorithm.

[0018] The present invention is further configured as follows: the spectrum analysis module includes a periodicity removal unit and a non-periodicity fitting unit;

[0019] The periodicity removal unit is used to perform multiple rounds of iterative Gaussian function fitting on the power spectrum, and then subtract the fitted Gaussian function from the power spectrum to obtain the remaining non-periodic signal;

[0020] The non-periodic fitting unit is used to perform non-periodic exponential fitting on the remaining non-periodic signals to obtain a spectrum slope.

[0021] The present invention also provides an attention detection method based on pupil size fluctuation characteristics, comprising the following steps:

[0022] S1. Collect pupil diameter data of the subject in a resting state using an eye tracker, with a sampling rate of no less than 200 Hz;

[0023] S2. Processing the pupil diameter data, deleting blinking data points, and supplementing missing data values ​​using linear interpolation to obtain interpolated pupil diameter data;

[0024] S3, converting the interpolated pupil diameter data, converting the time domain signal into a frequency domain signal using a Fourier transform algorithm, and calculating the power spectrum;

[0025] S4, performing a non-periodic exponential fitting on the power spectrum to generate a spectrum slope;

[0026] S5. Evaluate the attention level of the subject according to the spectrum slope.

[0027] The present invention is further configured such that the eye tracker in step S1 includes a glasses-type eye tracker and a desktop-type eye tracker.

[0028] The present invention is further configured as follows: the calculation formula of the Fourier transform algorithm in step S3 is:

[0029]

[0030] in, is the number of sample points, is a discrete-time signal, is a discrete frequency domain signal, k is frequency, n is time, e is a natural constant, and i is an imaginary unit.

[0031] The present invention is further configured as follows: the calculation formula of the power spectrum in step S3 is:

[0032]

[0033] Among them, x(f) represents the amplitude of the corresponding frequency, and its square value P(f) is the power spectrum. The power spectrum P(f) represents the change of the power of the frequency domain signal with frequency.

[0034] The present invention is further configured as follows: the calculation formula for the non-periodic exponential fitting in step S4 is:

[0035]

[0036] in, is the residual power at each frequency, F is the frequency, b is the spectrum shift, and x is the aperiodic index, i.e., the slope of the spectrum.

[0037] The present invention is further configured such that: the spectrum slope in step S5 is negatively correlated with the attention level, that is, the larger the spectrum slope, the lower the attention level, and the smaller the spectrum slope, the higher the attention level.

[0038] The present invention is further configured as follows: the method further comprises calculating a power ratio between two frequency bands, and using the power ratio instead of the spectrum slope to evaluate the attention level of the person being tested.

[0039] The present invention also provides an attention detection device based on pupil size fluctuation characteristics, comprising at least one processor; and a memory communicatively connected to at least one of the processors; wherein the memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement an attention detection method based on pupil size fluctuation characteristics.

[0040] The present invention also provides a computer-readable storage medium storing computer instructions, wherein the computer instructions are used to be executed by the computer to implement an attention detection method based on pupil size fluctuation characteristics.

[0041] The present invention also provides a computer program product, comprising a computer program, which, when executed by a processor, implements an attention detection method based on pupil size fluctuation characteristics.

[0042] In summary, the present invention has the following beneficial effects:

[0043] 1. This invention uses a glasses-type eye tracker with an infrared camera to collect pupil data. It does not require skin contact or the wearing of an electrode cap, thus avoiding the problems of conductive medium allergies and scalp discomfort in traditional EEG testing. The subject only needs to keep their eyes open and sit quietly. The operation process is non-invasive and is particularly suitable for children, patients with ADHD, and sensitive people.

[0044] 2. During the testing process, the present invention can complete pupil data collection in a resting state of only 5 minutes, without the need to complete specific tasks or maintain a fixed posture, which greatly shortens the testing time. At the same time, the glasses-type device supports Bluetooth data transmission, which simplifies the operation process and does not require professional installation and debugging. It can be flexibly applied in daily life or sports scenes;

[0045] 3. This invention utilizes pupil size fluctuations regulated by the autonomic nervous system, which is not controlled by subjective consciousness. This effectively avoids the drawbacks of deceptive responses in scale assessments. It uses a high-sampling rate (≥200Hz) infrared camera to obtain pupil data with millimeter-level precision. Combined with linear interpolation to fill in missing values, it significantly reduces interference from motion artifacts and environmental noise, thereby improving signal quality and signal-to-noise ratio.

[0046] 4. The quantitative indicator based on spectral slope analysis (the exponential fit slope of the non-periodic component) is directly related to the excitation-inhibition balance of the nervous system. The absolute value of the spectral slope is negatively correlated with the level of attention (the larger the absolute value of the slope, the lower the attention). This indicator has clear physiological significance and statistical significance, and is more sensitive and reliable than traditional TBR ratios or behavioral task performance assessments.

[0047] 5. The system of the present invention does not rely on EEG equipment or complex task environments and is suitable for a variety of scenarios such as schools, hospitals, and occupational health management. For example, it can be used for early screening of children with ADHD, or for real-time attention monitoring of special occupational groups such as drivers and security inspectors to prevent safety risks caused by lack of attention.

[0048] 6. The data processing module in the system of the present invention supports Fourier transform, Gaussian fitting and exponential fitting algorithms, and can further integrate machine learning models to optimize the accuracy of spectrum analysis. At the same time, the overall system architecture is flexible and can be connected to mobile terminals or cloud platforms to achieve remote monitoring and big data analysis, providing data support for personalized intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a data flow diagram of an attention detection system based on pupil size fluctuation characteristics in Example 1 of the present invention;

[0050] Figure 2This is a schematic diagram of an application process of an attention detection system based on pupil size fluctuation characteristics in Example 1 of the present invention;

[0051] Figure 3 1 is a flow chart of a method for attention detection based on pupil size fluctuation characteristics in Example 2 of the present invention;

[0052] Figure 4 This is a schematic diagram of the module structure of an attention detection system based on pupil size fluctuation characteristics in Example 1 of the present invention. DETAILED DESCRIPTION

[0053] The following is combined with Figure 1-Figure 4 The present invention is described in further detail.

[0054] Example 1: An attention detection system based on pupil size fluctuation characteristics, comprising a data acquisition module, a data processing module, a spectrum analysis module and a result evaluation module; in this embodiment, the data acquisition module is used to capture the eye image of the person being tested in real time, calculate the pupil diameter data and transmit it to the data processing module; the data processing module is used to receive the pupil diameter data transmitted by the data acquisition module for processing, delete the data points of blinking, supplement the missing data values ​​using linear interpolation, and then use the Fourier transform algorithm to convert the time domain signal into a frequency domain signal and transmit it to the spectrum analysis module; the spectrum analysis module is used to receive the frequency domain signal transmitted by the data processing module and calculate the power spectrum, and then generate the spectrum slope through non-periodic exponential fitting and transmit it to the result evaluation module; the result evaluation module is used to receive the spectrum slope transmitted by the spectrum analysis module to evaluate the attention level of the person being measured.

[0055] Preferably, in this embodiment, the data processing module includes a data preprocessing unit and a signal conversion unit; the data preprocessing unit is used to preprocess the pupil diameter data by deleting the data points of blinking and then using linear interpolation to supplement the missing data values ​​to obtain the interpolated pupil diameter data; the signal conversion unit is used to convert the interpolated pupil diameter data by converting the time domain signal into a frequency domain signal through a Fourier transform algorithm.

[0056] Preferably, in this embodiment, the spectrum analysis module includes a periodic removal unit and a non-periodic fitting unit; the periodic removal unit is used to perform multiple rounds of iterative Gaussian function fitting on the power spectrum, and then subtract the fitted Gaussian function from the power spectrum to obtain the remaining non-periodic signal; the non-periodic fitting unit is used to perform non-periodic exponential fitting on the remaining non-periodic signal to obtain the spectrum slope.

[0057] Example 2: A method for detecting attention based on pupil size fluctuation characteristics, comprising the following steps:

[0058] S1. Collect pupil diameter data of the subject in a resting state using an eye tracker, with a sampling rate of no less than 200 Hz.

[0059] In this embodiment, a quiet environment with constant brightness is selected as the test environment. During the test, the subject wears a glasses-type eye tracker (or a desktop eye tracker) and sits quietly with eyes open for 5 minutes in a resting state. The infrared camera on the eye tracker can capture the image of the eye in real time and calculate the diameter of the pupil. The data recording sampling rate is 200 Hz or above. The eye tracker then transmits the pupil diameter data to a computer via Bluetooth for further calculation.

[0060] S2. Process the pupil diameter data, delete the blinking data points, and use linear interpolation to supplement the missing data values ​​to obtain the interpolated pupil diameter data.

[0061] In this embodiment, 5 minutes of pupil size fluctuation data are preprocessed, blinking data points are deleted, and missing data values ​​are supplemented by linear interpolation.

[0062] S3. Convert the interpolated pupil diameter data, convert the time domain signal into a frequency domain signal through a Fourier transform algorithm, and calculate the power spectrum.

[0063] In this embodiment, the interpolated pupil diameter data is converted into a frequency domain signal using a Fourier transform algorithm. The calculation formula of the Fourier transform algorithm is:

[0064]

[0065] in, is the number of sample points, is a discrete-time signal, is a discrete frequency domain signal, k is frequency, n is time, e is a natural constant, and i is an imaginary unit.

[0066] The calculation formula of power spectrum is:

[0067]

[0068] Here, x(f) represents the amplitude of the corresponding frequency, and its square value P(f) is the power spectrum. The power spectrum P(f) represents the change of the power of the frequency domain signal with frequency. The power spectrum can be obtained by calculating the square of the amplitude spectrum.

[0069] S4. Perform a non-periodic exponential fitting on the power spectrum to generate a spectrum slope.

[0070] In this embodiment, multiple rounds of iterative Gaussian function fitting are performed on the spectrum, and then the fitted Gaussian function is subtracted from the spectrum. Then, a non-periodic exponential fitting is performed on the remaining signal. The calculation formula of the non-periodic exponential fitting is:

[0071]

[0072] in, is the residual power at each frequency, F is the frequency, b is the spectrum shift, and x is the aperiodic index, i.e., the slope of the spectrum.

[0073] S5. Assess the subject's attention level based on the spectrum slope.

[0074] In this embodiment, the attention level of the person being measured is evaluated based on the spectral slope of the pupil spectrum fluctuation, wherein the spectral slope is negatively correlated with the attention level, that is, the larger the spectral slope, the lower the attention level, and the smaller the spectral slope, the higher the attention level.

[0075] Preferably, in this embodiment, the method further comprises calculating a power ratio between two frequency bands, and using the power ratio instead of the spectrum slope to evaluate the attention level of the subject.

[0076] Example 3: An attention detection device based on pupil size fluctuation characteristics, comprising at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the processor, the instructions being used to be executed by the processor to implement an attention detection method based on pupil size fluctuation characteristics, the method comprising: collecting pupil diameter data of a subject in a resting state through an eye tracker, with a sampling rate of not less than 200 Hz; processing the pupil diameter data, deleting data points of blinking, and supplementing missing data values ​​using linear interpolation to obtain interpolated pupil diameter data; converting the interpolated pupil diameter data, converting the time domain signal into a frequency domain signal through a Fourier transform algorithm, and calculating a power spectrum; performing non-periodic exponential fitting on the power spectrum to generate a spectral slope; and evaluating the subject's attention level based on the spectral slope.

[0077] Example 4: A computer-readable storage medium stores computer instructions, which are used to be executed by a computer to implement an attention detection method based on pupil size fluctuation characteristics, the method comprising: collecting pupil diameter data of a subject in a resting state through an eye tracker, with a sampling rate of not less than 200 Hz; processing the pupil diameter data, deleting data points of blinking, and supplementing missing data values ​​using linear interpolation to obtain interpolated pupil diameter data; converting the interpolated pupil diameter data, converting the time domain signal into a frequency domain signal through a Fourier transform algorithm, and calculating a power spectrum; performing non-periodic exponential fitting on the power spectrum to generate a spectral slope; and evaluating the subject's attention level based on the spectral slope.

[0078] Example 5: A computer program product, comprising a computer program, which, when executed by a processor, implements an attention detection method based on pupil size fluctuation characteristics, the method comprising: collecting pupil diameter data of a subject in a resting state through an eye tracker, with a sampling rate of not less than 200 Hz; processing the pupil diameter data, deleting data points of blinking, and supplementing missing data values ​​using linear interpolation to obtain interpolated pupil diameter data; converting the interpolated pupil diameter data, converting the time domain signal into a frequency domain signal through a Fourier transform algorithm, and calculating a power spectrum; performing non-periodic exponential fitting on the power spectrum to generate a spectral slope; and evaluating the subject's attention level based on the spectral slope.

[0079] This specific embodiment is merely an explanation of the present invention and is not intended to limit the present invention. After reading this specification, those skilled in the art may make non-creative modifications to this embodiment as needed. However, as long as such modifications are within the scope of the claims of the present invention, they are protected by patent law.

Claims

1. An attention detection system based on pupil size fluctuation characteristics, characterized by: It includes data acquisition module, data processing module, spectrum analysis module and result evaluation module; The data acquisition module is used to capture the eye image of the subject in real time, calculate the pupil diameter data and transmit it to the data processing module; The data processing module is used to receive and process pupil diameter data transmitted by the data acquisition module, delete blinking data points, supplement missing data values ​​using linear interpolation, and then use Fourier transform algorithm to convert time domain signals into frequency domain signals and transmit them to the spectrum analysis module; The spectrum analysis module is used to receive the frequency domain signal transmitted by the data processing module and calculate the power spectrum, and then generate the spectrum slope through non-periodic exponential fitting and transmit it to the result evaluation module; The result evaluation module is used to receive the spectrum slope transmitted by the spectrum analysis module to evaluate the attention level of the measured person; The data processing module includes a data preprocessing unit and a signal conversion unit; The data preprocessing unit is used to preprocess the pupil diameter data by deleting the data points of blinking and then supplementing the missing data values ​​by linear interpolation to obtain the pupil diameter data after interpolation; The signal conversion unit is used to convert the interpolated pupil diameter data, and convert the time domain signal into a frequency domain signal through a Fourier transform algorithm; The spectrum analysis module includes a periodicity removal unit and a non-periodicity fitting unit; The periodicity removal unit is used to perform multiple rounds of iterative Gaussian function fitting on the power spectrum, and then subtract the fitted Gaussian function from the power spectrum to obtain the remaining non-periodic signal; The non-periodic fitting unit is used to perform non-periodic exponential fitting on the remaining non-periodic signals to obtain a spectrum slope.

2. A method for detecting attention based on pupil size fluctuation characteristics, applied to the attention detection system based on pupil size fluctuation characteristics of claim 1, characterized in that: The following steps are involved: S1. Collect pupil diameter data of the subject in a resting state using an eye tracker, with a sampling rate of no less than 200 Hz; S2. Processing the pupil diameter data, deleting blinking data points, and supplementing missing data values ​​using linear interpolation to obtain interpolated pupil diameter data; S3, converting the interpolated pupil diameter data, converting the time domain signal into a frequency domain signal using a Fourier transform algorithm, and calculating the power spectrum; S4, performing a non-periodic exponential fitting on the power spectrum to generate a spectrum slope; S5. Evaluate the attention level of the subject according to the spectrum slope.

3. The method for detecting attention based on pupil size fluctuation characteristics according to claim 2, characterized in that: The eye tracker in step S1 includes a glasses-type eye tracker and a desktop-type eye tracker.

4. The method for detecting attention based on pupil size fluctuation characteristics according to claim 2, characterized in that: The spectrum slope in step S5 is negatively correlated with the attention level, that is, the larger the spectrum slope, the lower the attention level, and the smaller the spectrum slope, the higher the attention level.

5. The method for detecting attention based on pupil size fluctuation characteristics according to claim 2, characterized in that: The method further comprises calculating a power ratio between two frequency bands, and using the power ratio instead of the spectrum slope to evaluate the attention level of the subject.

6. An attention detection device based on pupil size fluctuation characteristics, characterized by: It includes at least one processor; and a memory communicatively connected to at least one of the processors; wherein the memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement an attention detection method based on pupil size fluctuation characteristics as described in any one of claims 2-5.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the attention detection method based on pupil size fluctuation characteristics according to any one of claims 2 to 5.

8. A computer program product, characterized in that: The invention comprises a computer program, which, when executed by a processor, implements an attention detection method based on pupil size fluctuation characteristics according to any one of claims 2 to 5.

Images