An AI emotion recognition mental health screening system

Abstract

The present invention discloses an AI emotion recognition mental health screening system, which relates to the field of image recognition technology, including a processing chip, a control register, a photosensitive multi-element converter, a first frame memory unit, a second frame memory unit, a frame difference capture unit, a frame difference summator, an externally electrically connected signal output unit, the signal output ends of the first frame memory unit and the second frame memory unit are electrically connected to the frame difference capture unit, the signal output end of the frame difference capture unit is electrically connected to the frame difference summator, and the signal output end of the frame difference summator is electrically connected to the signal output unit. The present invention is simple to use, intelligent learning, multi-dimensional analysis and widely used, and can identify detection areas that cannot be identified by face technology, as well as the recognition of facial orientation in areas that cannot be detected by face technology, and can also identify faces with partial occlusion and faces with light above 100LUX, and is not affected by expression, age, or race.

Description

Technical Field

[0001] The present invention belongs to the technical field of image recognition, and in particular relates to an AI emotion recognition mental health screening system. Background Art

[0002] Currently, the main recognition technologies include face recognition technology and Neuros artificial intelligence. These recognition technologies have certain shortcomings. For example, face recognition technology obtains human emotions through micro-expressions, which is easy to disguise and has strict requirements on the face detection area; while Neuros artificial intelligence is based on blood spectrum optical technology and imaging technology, which has very high requirements on light. These two methods cannot effectively identify once the face is blocked. Therefore, according to the needs, the present invention proposes an AI emotion recognition mental health screening system. Summary of the invention

[0003] The purpose of the present invention is to solve the above shortcomings in the prior art and propose an AI emotion recognition mental health screening system. The micro-vibration of a person's head and neck is collected and the person's emotions are calculated, which can be simple to use, intelligently learn, multi-dimensionally analyzed and widely used.

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

[0005] An AI emotion recognition mental health screening system is designed, characterized in that it includes a processing chip, a control register, a photosensitive multi-converter, a first frame memory unit, a second frame memory unit, a frame difference capture unit, a frame difference summer, and an externally electrically connected signal output unit integrated in the processing chip, the input and output ends of the control register are electrically connected to the photosensitive multi-converter and the frame difference summer, respectively; the signal output end of the photosensitive multi-converter is electrically connected to the signal input end of the first frame memory unit and the second frame memory unit, the optical image is converted into an electrical signal in the photosensitive multi-converter, and the electrical signal is read out frame by frame to the first frame memory in sequence. The first frame memory unit and the second frame memory unit are in the second frame memory unit; the signal output ends of the first frame memory unit and the second frame memory unit are electrically connected to the frame difference capture unit, which is used for frame difference capture of signal synchronization of the first frame memory unit and the second frame memory unit; the signal output end of the frame difference capture unit is electrically connected to the frame difference summerizer, which is used for the frame difference capture unit to feed the signal to the frame difference summerizer, and the signal undergoes frame difference accumulation in the frame difference summerizer to form an image frame difference, thereby obtaining an external halo of the optical image; the signal output end of the frame difference summerizer is electrically connected to the signal output unit, and the image frame difference signal obtained after the signal is summed by the frame difference summerizer is sent to the external signal output unit.

[0006] Furthermore, frame difference accumulation occurs in the frame difference summer to form image frame differences including amplitude frame differences and frequency frame differences.

[0007] Furthermore, the external halo of the optical image obtained in the frame difference summer is the external vibration parameter of the head in the optical image, including the maximum frequency of micro-vibration on the straight line and the average amplitude of micro-vibration on the straight line, and the external vibration parameter constitutes the emotional state of human energy change.

[0008] Furthermore, the emotional state includes an aggressiveness value, a stress value, an anxiety value, a suspiciousness value, a balance value, a charm value, a vitality value, a self-regulation value, a repression value, a neuroticism value, a depression value, a happiness value, an extroversion value, and a stability value.

[0009] Compared with the prior art, the AI ​​emotion recognition mental health screening system proposed by the present invention has the following beneficial effects: it collects the micro-vibration of a person's head and neck and calculates the person's emotions, and can be simple to use, intelligently learn, multi-dimensionally analyzed and widely used. Specifically:

[0010] (1) Easy to use, non-contact analysis and detection, only 3 to 60 seconds of video acquisition is required to quickly detect and produce results.

[0011] (2) Intelligent learning uses the Q-Learning algorithm in machine learning to make the test value more and more accurate.

[0012] (3) Multidimensional analysis, multidimensional psychological and physiological indicators such as fatigue, attention, psychological state and other psychological test results.

[0013] (4) It has a wide range of applications and can be used in different scenarios, such as psychological and physiological health screening, security, polygraph, marketing, education, interviews, etc.

[0014] (5) The present invention can also identify detection areas that cannot be identified by human face technology, and recognize the facial orientation of areas that cannot be detected by human face technology.

[0015] (6) The present invention can also recognize faces with partial occlusion and faces with light intensity above 100 LUX, and is not affected by expression, age, or race. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0017] Figure 1 This is a system block diagram of the AI ​​emotion recognition mental health screening system of the present invention;

[0018] Figure 2 This is a flowchart of the implementation of the AI ​​emotion recognition mental health screening system of the present invention;

[0019] Marked in the figure are: 1, processing chip; 2, control register; 3, photosensitive multi-element converter; 41, first frame memory unit; 42, second frame memory unit; 5, frame difference capture unit; 6, frame difference summer; 7, signal output unit. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention; it is obvious that the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments, and all other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making creative work are within the scope of protection of the present invention.

[0021] In the description of the present invention, it should be noted that the positions or positional relationships indicated by the terms "input end", "output end", etc. are based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the devices or elements referred to must have a specific position, be constructed and operated in a specific position, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0022] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "provided with", "connected", etc. should be understood in a broad sense. For example, "connected" 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 it can be indirectly connected 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 present invention can be understood according to specific circumstances.

[0023] The structural features of the present invention are now described in detail with reference to the accompanying drawings.

[0024] See also Figure 1, an AI emotion recognition mental health screening system, which is composed of acquiring a sequence frame of a target image, subtracting two frames and subsequent processing of frame difference, comprises a processing chip 1, a control register 2, a photosensitive multi-converter 3, a first frame memory unit 41, a second frame memory unit 42, a frame difference capture unit 5, a frame difference summator 6, and an externally electrically connected signal output unit 7, wherein the input and output ends of the control register 2 are electrically connected to the photosensitive multi-converter 3 and the frame difference summator 6 respectively. The signal output end of the photosensitive multi-converter 3 is electrically connected to the signal input end of the first frame memory unit 41 and the second frame memory unit 42, the optical image is converted into an electrical signal in the photosensitive multi-converter 3, and the electrical signal is sequentially read out frame by frame into the first frame memory unit 41 and the second frame memory unit 42; the signal output end of the first frame memory unit 41 and the second frame memory unit 42 are electrically connected to the frame difference capture unit 5, for frame difference capture of the first frame memory unit 41 and the second frame memory unit 42 signals being synchronized. The signal output end of the frame difference capture unit 5 is electrically connected to the frame difference summator 6, so that the frame difference capture unit 5 feeds the signal to the frame difference summator 6, and the signal accumulates frame differences in the frame difference summator 6 to form an image frame difference, thereby obtaining an external halo of the optical image. The signal output end of the frame difference summator 6 is electrically connected to the signal output unit 7, and the image frame difference signal obtained after the frame difference summation of the signal by the frame difference summator 6 is sent to the external signal output unit 7.

[0025] The formula for frame difference accumulation in the frame difference summer 6 is: , formula (1)

[0026] in, is the real signal of the i-th element in the k-th frame; is the real signal of the i-th element in the k+1-th frame; M is the number of frames being processed in the frame sequence, which depends on the frame difference accumulation time and the performance speed of the device used; L is the exponential coefficient of the display process 3 or 4; N is the average coefficient; D is the threshold of the image noise. Frame difference accumulation occurs in the frame difference summer 6, and the image frame difference includes amplitude frame difference and frequency frame difference. Among them, the difference formula of the amplitude frame difference is: ,Formula (2)

[0027] in: is the signal at point x, y in the i-th frame, is the signal at point x, y in the i+1th frame; is the number of frames for accumulating the amplitude component of the vibration image.

[0028] The frequency frame difference formula is: ,Formula (3)

[0029] in: The difference between the frames of the image at point i is 1 if there is a difference, and 0 if there is no difference. is the number of frames for accumulating the amplitude component of the vibration image.

[0030] The external halo of the optical image obtained in the frame difference summer 6 is the external vibration parameter of the head in the optical image, including the maximum frequency of the micro-vibration on the straight line and the average amplitude of the micro-vibration on the straight line. The external vibration parameter constitutes the emotional state of the human body energy change, and its formula is: ,Formula (4)

[0031] in, For a certain emotional state, is the frequency of change in time of work done by a person's head in the form of micro-vibration, The spatially distributed displacement of a person's head doing work in the form of micro-vibrations; : Corresponding relationship.

[0032] Emotional states include aggression value, stress value, anxiety value, suspicion value, balance value, charm value, vitality value, self-regulation value, repression value, neuroticism value, depression value, happiness value, extroversion value and stability value. The model algorithm equation for the aggression value is: ,in, is the maximum frequency in the frequency distribution density histogram; F i is the i-th frequency in the frequency distribution density histogram; is the input processing vibration frequency; n is The count of intra-frame and inter-frame differences greater than the threshold; the aggressiveness is determined by the vibration frequency histogram, which reflects the maximum value of the frequency distribution density and the standard deviation of the person's vibration.

[0033] The model algorithm equation for the pressure value is:

[0034] ,

[0035] in, is the total amplitude of the i-th line vibration frequency component on the left side of the object; is the total amplitude of the frequency component of the ith line vibration on the right side of the object; for ~ The maximum value between is the maximum frequency of the vibration frequency component of the i-th row on the left side of the object; is the maximum frequency of the vibration frequency component of the i-th row on the right side of the object, for ~ The maximum value between is the number of lines occupied by the image; the degree of asymmetry of the external frequency vibration reflects the level of pressure.

[0036] The model algorithm equation of anxiety value is:

[0037] ,

[0038] in, is the spectral power of the vibration frequency distribution; It is the maximum frequency in the vibration frequency distribution spectrum; high-density high-frequency vibrations reflect high anxiety levels.

[0039] The model algorithm equation for suspicious values ​​is:

[0040] ,

[0041] Among them, E1 is the aggressiveness level; E2 is the stress level; E3 is the anxiety level; aggressiveness, stress, and anxiety reflect the suspiciousness level.

[0042] The equation for the equilibrium value is:

[0043] ,

[0044] in, The sum is calculated for the emotion parameter variability; the ratio of the standard deviation and the mathematical expectation reflects the emotion parameter variability.

[0045] The model algorithm equation of charm value is:

[0046] ,

[0047] in is the difference between the left and right amplitude averages of the vibration image amplitude component of each line; is the difference between the maximum frequency values ​​on the left and right sides of the vibration image amplitude component for each line; is the number of frames in the processing process; the symmetry of head micro-movement reflects the charm value.

[0048] The model algorithm equation of vitality value is:

[0049] ,

[0050] Where M is the maximum value of the count on the frequency histogram; The standard deviation of the vibration frequency, calculated from the frequency histogram; The maximum value of the input frequency for the vibration image; the difference between the maximum density and the standard deviation of the vibration frequency.

[0051] The model algorithm equation for the self-regulating value is:

[0052] ,

[0053] Among them, E5 is the average value of the balance parameter during the measurement process; dE5 is the variation range of the balance parameter; E6 is the average value of the charm value parameter during the measurement process; dE6 is the variation range of the charm value parameter; the self-regulation level reflects the stability of positive emotions (vitality value and charm value).

[0054] The model algorithm equation for the inhibition value is:

[0055] ,

[0056] in, is the vibration change frequency; is the average period of vibration frequency change; T: vibration period measurement; inhibition level comes from human reaction time to stimulation.

[0057] The model algorithm equation for neuroticism is:

[0058] ,

[0059] in, is the standard deviation of the E9 (inhibition) values; the standard deviation of the inhibition parameter reflects the level of neuroticism.

[0060] The model algorithm equation for depression value is:

[0061] ,

[0062] in, is the standard deviation of the vibration frequency in the frequency histogram; is the average value of the vibration frequency in the frequency histogram; the standard deviation and average value of the vibration frequency in the frequency histogram reflect the depression level.

[0063] The model algorithm equation of happiness value is:

[0064] ,

[0065] Among them, I is the information effectiveness of the psychological and physiological state; E is the energy reduction characteristic of the psychological and physiological state; dI is the change in the information effectiveness of the psychological and physiological state; dE is the change in the energy reduction characteristic of the psychological and physiological state; information efficiency and energy consumption reflect the level of happiness.

[0066] The model algorithm equation for the extraversion value is:

[0067] ,

[0068] in, is the Pearson correlation coefficient between information efficiency (I) and energy expenditure (E); the correlation between information efficiency and energy expenditure reflects the level of extraversion.

[0069] The model algorithm equation for the stability value is:

[0070] ,

[0071] Among them, K is the normalization coefficient of the frequency histogram, is the normal distribution density;

[0072] ; ; The similarity between the frequency histogram and the normal distribution law reflects the level of stability.

[0073] Reference Figure 2 Specifically, the optical image is converted into an electrical signal in the photosensitive multiplexer of module 3, and the electrical signal is read out frame by frame in turn into the first and second frame memory units 41 and 42, and synchronized to the frame difference capture unit 5, and the signal is fed to the frame difference summer 6, where frame difference accumulation occurs and is sent to the signal output unit 7. The photosensitive multiplexer 3 converts the head vibration in the optical image into motion parameters of the head points. There are differences in optical contrast between points. The frame difference is accumulated at each point through the frame difference summer 6 to form vibration imaging. The optical contrast difference includes the change of the amplitude of the signal at each point in space and the change of the frequency of the signal at each point in space.

[0074] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, technical students in the field can still modify the technical solutions recorded in the aforementioned embodiments, or replace some of the technical features therein with equivalents. 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. An AI emotion recognition mental health screening system, characterized in that: The invention comprises a processing chip (1), a control register (2), a photosensitive multi-element converter (3), a first frame memory unit (41), a second frame memory unit (42), a frame difference capture unit (5), a frame difference summator (6), and an externally electrically connected signal output unit (7), wherein the input and output ends of the control register (2) are electrically connected to the photosensitive multi-element converter (3) and the frame difference summator (6) respectively; The signal output end of the photosensitive multi-element converter (3) is electrically connected to the signal input ends of the first frame memory unit (41) and the second frame memory unit (42); the optical image is converted into an electrical signal in the photosensitive multi-element converter (3); and the electrical signal is read out frame by frame in sequence into the first frame memory unit (41) and the second frame memory unit (42); The signal output ends of the first frame memory unit (41) and the second frame memory unit (42) are electrically connected to a frame difference capture unit (5) for frame difference capture of signal synchronization between the first frame memory unit (41) and the second frame memory unit (42); The signal output end of the frame difference capture unit (5) is electrically connected to the frame difference summator (6), so that the frame difference capture unit (5) feeds the signal to the frame difference summator (6), and the signal is accumulated through the frame difference summator (6) to form an image frame difference, thereby obtaining an external halo of the optical image; The signal output end of the frame difference summator (6) is electrically connected to a signal output unit (7), and an image frame difference signal obtained after the frame difference summation of the signal by the frame difference summator (6) is sent to an external signal output unit (7); The external halo of the optical image obtained in the frame difference summer (6) is the external vibration parameter of the head in the optical image, including the maximum frequency of the micro-vibration on the straight line and the average amplitude of the micro-vibration on the straight line. The external vibration parameter constitutes the emotional state of the human body energy change, and its formula is: ,Formula (4) in, For a certain emotional state, The frequency is the time variation of a person's head doing work in the form of micro-vibration. The spatially distributed displacement of a person's head doing work in the form of micro-vibrations; For the corresponding relationship; The emotional state includes an aggression value, a stress value, an anxiety value, a suspicious value, a balance value, a charm value, a vitality value, a self-regulation value, a repression value, a neuroticism value, a depression value, a happiness value, an extroversion value, and a stability value; Among them, the model algorithm equation of the aggressiveness value is: , Formula (5) in, is the maximum frequency in the frequency distribution density histogram; is the average frequency in the frequency distribution density histogram; is the input processing vibration frequency; n is The count of intra-frame and inter-frame differences greater than the threshold, is the number of frames that accumulate the amplitude component of the vibration image, and the aggressiveness is determined by the vibration frequency histogram, which reflects the maximum value of the frequency distribution density and the standard deviation of the person's vibration; is the standard deviation of the vibration frequency in the frequency histogram; The model algorithm equation for the pressure value is: , Formula (6) in, is the total amplitude of the i-th line vibration frequency component on the left side of the object; is the total amplitude of the frequency component of the ith line vibration on the right side of the object; for ~ The maximum value between is the maximum frequency of the vibration frequency component of the i-th row on the left side of the object; is the maximum frequency of the vibration frequency component of the i-th row on the right side of the object, for ~ The maximum value between is the number of rows occupied by the image; The model algorithm equation for the anxiety value is: , Formula (7) in, is the spectral power of the vibration frequency distribution; is the maximum frequency in the vibration frequency distribution spectrum; The model algorithm equation for suspicious values ​​is: , Formula (8) The equation for the equilibrium value is: , Formula (9) in, Calculate the sum for the emotion parameter variability; The model algorithm equation of charm value is: , Formula (10) in, is the difference between the left and right amplitude averages of the vibration image amplitude component of each line; is the difference between the maximum frequency values ​​on the left and right sides of the vibration image amplitude component for each line; is the number of frames in the processing process; The model algorithm equation of vitality value is: , Formula (11) Where M is the maximum value of the count on the frequency histogram; is the standard deviation of the vibration frequency in the frequency histogram, calculated by the frequency histogram; Enter the maximum value of the frequency for the vibration image; The model algorithm equation for the self-regulating value is: , Formula (12) Among them, dE5 is the range of change of the balance parameter; dE6 is the range of change of the charm value parameter; The model algorithm equation of the suppression value is: , Formula (13) in, is the vibration change frequency; is the average period of vibration frequency change; T is the vibration period measurement; The model algorithm equation for neuroticism is: , Formula (14) in, is the standard deviation of the E9 suppression value; The model algorithm equation for depression value is: , Formula (15) in, is the standard deviation of the vibration frequency in the frequency histogram; is the average value of the vibration frequency in the frequency histogram; The model algorithm equation of happiness value is: , Formula (16) Where I is the information validity of the psychological and physiological state; E is the energy reduction characteristic of the psychological and physiological state; dI is the change in the information validity of the psychological and physiological state; dE is the change in the energy reduction characteristic of the psychological and physiological state; The model algorithm equation for the extraversion value is: , Formula (17) in, is the Pearson correlation coefficient between information efficiency I and energy consumption E; The model algorithm equation for the stability value is: , Formula (18) Among them, K is the normalization coefficient of the frequency histogram, is the normal distribution density; ； ； Where M is the maximum value of the count on the frequency histogram, is the standard deviation of the vibration frequency in the frequency histogram.

2. The AI ​​emotion recognition mental health screening system according to claim 1, characterized in that: The formula for frame difference accumulation in the frame difference summer (6) is: , Formula (1) in, is the real signal of the i-th element in the k-th frame; is the real signal of the i-th element in the k+1-th frame; M is the number of frames being processed in the frame sequence, which depends on the frame difference accumulation time and the performance speed of the device used; L is the exponential coefficient of the display process 3 or 4; N is the average coefficient; D is the threshold of image noise.

3. The AI ​​emotion recognition mental health screening system according to claim 1 or 2, characterized in that: The frame difference summer (6) accumulates the frame difference to form an image frame difference including an amplitude frame difference and a frequency frame difference; The difference formula of the amplitude frame difference is: ,Formula (2) in: The signal at point (x, y) in the i-th frame, The signal at the point (x, y) in the i+1th frame; The frequency frame difference calculation formula is: ,Formula (3) in: is the maximum frequency in the frequency distribution density histogram, is the inter-frame difference of the image at the i-th point, 1 if there is a difference, and 0 if there is no difference.

Images