Emotion recognition system and method for eye tracker
Technical Field
The invention belongs to the technical field of eye movement instruments, and particularly relates to an emotion recognition system and method for an eye movement instrument.
Background
The eye tracker is an important instrument for basic research of psychology. The eye tracker is used for recording eye movement track characteristics of people when processing visual information, and is widely used for researches in the fields of attention, visual perception, reading and the like. The temporal and spatial features of eye movement are physiological and behavioral manifestations in the process of visual information extraction, which have a direct or indirect relationship with human mental activities, and are the reason why many psychologists are dedicated to eye movement research.
Most of existing equipment for emotion data analysis by combining with an eye tracker only records pupil movement data and is combined to be applied to autism, AR scene application and VR scene application.
For example, chinese patent document CN111339878A discloses a modified real-time emotion recognition method and system based on eye movement data, which collects eye movement data in real time in VR environment, i.e. can accurately and efficiently acquire rich eye movement information (eye movement time, eye movement direction, eye movement distance, gaze time, gaze frequency, pupil diameter size, and blink frequency) by tracking with an eye tracker, analyze the emotional state of the subject from the eye movement angle, and process facial expression feature vectors and eye movement feature vectors using SVR.
For example, chinese patent documents CN105868694A and CN109803109A disclose an emotion recognition scheme based on facial expression and eyeball motion, which utilizes an eye tracker to capture a focal position focused by eyes of a user, utilizes a facial expression recognition system to collect physiological signals of 8 collection positions of a human face, analyzes a muscle tension degree of each collection position according to the physiological signals of each collection position, and then combines the muscle tension degrees of the 8 collection positions to analyze and determine a current facial expression of the user through an analysis algorithm.
For example, chinese patent document CN110931111A discloses an autism assistance intervention system and method based on virtual reality and multi-modal information, which uses a built-in eye tracker to obtain eye movement data, calculates the direction and focus of eye movement, uses an extraluminal camera to photograph the lower half of the face to obtain muscle movement data of the chin and both cheeks, and performs emotion recognition through a speech feature extraction unit to obtain emotion scores of user interaction languages.
For example, chinese patent document CN111695442A discloses an online intelligent assistance system based on multi-mode fusion, in which an eye tracker detects the eye jump process of a user, extracts corresponding features from data collected by the eye tracker through a convolutional neural network, and determines the gas quality type through an SVM.
However, there is no emotion recognition solution for an eye tracker that not only tracks the eye movement state, but also performs comprehensive analysis in combination with data such as eyelid and blink states, and compares the eye movement state with the data of each time to obtain weighted emotion data reference.
Disclosure of Invention
In view of the technical problems in the prior art, an object of the present invention is to provide an emotion recognition system and method for an eye tracker, which can track the state of eye movement, perform comprehensive analysis by combining data such as eyelid and blink states, compare the data of each time to obtain weighted emotion data reference, and can be integrated into a solution of the eye tracker.
In order to achieve the purpose, the invention provides the following technical scheme: an emotion recognition system for an eye tracker, which comprises a picture playing system, an eye movement monitoring system and an emotion analysis system, the picture playing system is used for playing a user watching video picture, the eye movement monitoring system comprises a light supplementing system which is used as a light source to realize face light supplementation, a pupil tracking system which is combined with a camera device to track the position of a pupil and store tracking data, a pupil size monitoring system which monitors the size of the pupil, a blink monitoring system which monitors the blink state and an eyelid relaxation monitoring system which monitors the eyelid relaxation degree, the emotion analysis system comprises an eye movement information base system, an eye movement information comparison system for comprehensively comparing the data in the eye movement information base system and the eye movement monitoring system, and further comprises an emotion information base system and an emotion information comparison system for comprehensively comparing the data in the emotion information base system and the eye movement monitoring system.
Preferably, the pupil tracking system includes a track imaging system for recording a pupil movement track, an eye jump frequency monitoring system for monitoring a pupil jump frequency, an eye jump distance monitoring system for monitoring and recording a pupil jump distance in combination with the eye jump frequency monitoring system, and a displacement prediction system for predicting the displacement of the pupil before opening the eye in combination with the blink monitoring system and the eyelid relaxation monitoring system.
Preferably, the blink monitoring system is used for blink detection, wherein the blink monitoring system comprises a blink speed monitoring system for monitoring blink speed, a blink frequency monitoring system for monitoring blink frequency and a blink time recording system for monitoring blink time, the blink speed monitoring system and the blink frequency monitoring system are combined to record blink speed data, the blink speed and the blink frequency are measured in the same time length, the blink speed detection is averaged to be a, the blink speed exceeding A is recorded to be Aa, and the blink speed lower than A is recorded to be Ab. The blink frequency detection is averaged to obtain a mean value V, and blink frequencies above V are recorded as Va, and blink frequencies below V are recorded as Vb. Generally speaking, a fast blink speed indicates that the user is in a good active state, the follow-up picture is compact, the tension is high, the attention is relatively concentrated, and a fast blink frequency indicates that the user has high interest in the picture and the follow-up is good.
Preferably, the eyelid laxity monitoring system is used for eyelid laxity detection, wherein detection results of the eyelid laxity monitoring system are distinguished by the degree of coverage of the iris, including an uncovered state B, a partial coverage Ba, a half coverage Bb and a majority coverage Bc. By recording the degree of coverage of the iris, the eyelid laxity of the user can be judged. The greater the coverage the higher the sag.
According to another aspect of the present invention, the present invention further provides an emotion recognition method for an eye tracker, comprising the steps of:
s1, starting a picture playing system and starting a light supplementing system to supplement light on the surface of the human face, and acquiring eye movement information within a certain time through an eye movement monitoring system within a light ray range meeting the identification requirement;
s2, comparing the eye movement information obtained in the S1 with the eye movement information database system and the eye movement information comparison system to obtain basic comparison data;
and S3, comparing the basic comparison data obtained in the S2 with the emotion information database system and the emotion information comparison system, performing weight analysis with different emotion data, and obtaining analysis results with different weights.
Preferably, the data comparison content in S2 includes pupil trajectory comparison, eye jump frequency and eye jump distance in combination with a pupil tracking system, and further includes pupil size comparison in combination with a pupil size monitoring system, blink speed comparison and blink frequency comparison in combination with a blink monitoring system, and eyelid sag comparison in combination with an eyelid sag monitoring system.
According to the emotion recognition system and method for the eye movement instrument, due to the configuration of the light supplementing system, light can be supplemented to the face under the dark light condition, and the eye movement monitoring system can accurately capture eye movement data.
According to the emotion recognition system and method for the eye movement instrument, the pupil tracking system records eye movement data, and the displacement prediction system can predict and capture the approximate position of a pupil in advance in a blinking state, so that the pupil position can be conveniently and quickly positioned when an eyelid is opened; the eye movement information base system in the emotion analysis system is combined with the eye movement information comparison system to compare the eye movement information, can store the current eye movement information and facilitate comparison in the follow-up eye movement information collection process, has the function of increasing comparison information, and realizes data learning.
According to the emotion recognition system and method for the eye tracker, the emotion information base system is arranged to be combined with the emotion information comparison system, so that emotion weighting of current eye movement information can be combined, and a weight proportion is provided for reference.
According to the emotion recognition system and method for the eye tracker, the technical scheme that the emotion recognition system and the eye tracker are integrated into a whole can be realized no matter the eye tracker is an external eye tracker or an internal eye tracker, and the intelligence of the eye tracker is improved.
Drawings
FIG. 1 is a schematic block diagram illustrating an emotion recognition system for an eye tracker in accordance with an embodiment of the present invention;
FIG. 2 is a schematic block diagram illustrating a pupil tracking system for an emotion recognition system for an eye tracker in accordance with an embodiment of the present invention;
fig. 3 is a schematic block diagram illustrating a blink monitoring system for an emotion recognition system of an eye tracker according to an embodiment of the present invention.
The reference numbers in the figures are respectively: the system comprises a picture playing system 1, an eye movement monitoring system 2, an emotion analysis system 3, a light supplementing system 4, a pupil tracking system 5, a track imaging system 501, an eye jump frequency monitoring system 502, an eye jump distance monitoring system 503, a displacement prediction system 504, a pupil size monitoring system 6, a blink monitoring system 7, a blink speed monitoring system 701, a blink frequency monitoring system 702, a blink time recording system 703, an eyelid looseness monitoring system 8, an eye movement information base system 9, an eye movement information comparison system 10, an emotion information base system 11 and an emotion information comparison system 12.
Detailed Description
The present invention will now be described more fully hereinafter with reference to the accompanying drawings, in which embodiments of the invention are shown and described, and it is to be understood by those skilled in the art that the foregoing detailed description is illustrative, and that the invention is not limited to the specific embodiments shown.
Unless otherwise indicated, all references to up, down, left, right, front, back, inner and outer directions herein are to be interpreted as referring to up, down, left, right, front, back, inner and outer directions in the drawings to which the invention is directed.
FIG. 1 is a schematic block diagram of an emotion recognition system for an eye tracker; FIG. 2 is a schematic block diagram of a pupil tracking system; fig. 3 is a schematic block diagram of a blink monitoring system.
As shown in fig. 1, in an embodiment of the present invention, an emotion recognition system for an eye tracker is provided, which includes a picture playing system 1, an eye monitoring system 2, and an emotion analysis system 3, where the picture playing system 1 is configured to play a video picture watched by a user, and guide the eyeball of the user to watch the picture content interested by the user through the picture playing system. Eye movement monitoring system 2 includes and realizes facial light supplemental light system 4 as the light source, through the configuration of light supplemental light system 4, can guarantee to carry out the light supplement to face under the dim light condition, and to the occasion of seeing the shadow, light often appears inadequately, and receives the influence of broadcast picture content and the change in luminance is big, and the accuracy that detecting system detected is ensured to the light supplement through this light supplemental light system 4 this moment. The eye movement monitoring system 2 further comprises a pupil tracking system 5, the pupil tracking system 5 captures eye movement data, and can track the position of a pupil by combining with an image pickup device and store the tracking data in a memory, the eye movement monitoring system 2 further comprises a pupil size monitoring system 6 for monitoring the size of the pupil, the pupil size monitoring system 6 detects pupil size data and data of pupil size change, in a preferred embodiment, the pupil size monitoring system 6 also detects the size and the changed data of an iris and stores the data in the memory, and the eye movement monitoring system 2 further comprises a blink monitoring system 7 for monitoring a blink state and an eyelid looseness monitoring system 8 for monitoring the eyelid looseness degree. The blink monitoring system detects a blink state. The eyelid laxity monitoring system 8 is used for eyelid laxity detection, and the detection result is distinguished by the degree of coverage of the iris, including an uncovered state B, a partial coverage Ba, a half coverage Bb and a large coverage Bc. Preferably, the eyelid laxity monitoring system is used for eyelid laxity detection, wherein detection results of the eyelid laxity monitoring system are distinguished by the degree of coverage of the iris, including an uncovered state B, a partial coverage Ba, a half coverage Bb and a majority coverage Bc. By recording the degree of coverage of the iris, the eyelid laxity of the user can be judged. The greater the coverage the higher the sag.
In an embodiment of the present invention, the emotion analyzing system 3 of the emotion recognition system includes an eye movement information base system 9, and an eye movement information comparing system 10 for performing comprehensive comparison by combining data in the eye movement information base system 9 and the eye movement monitoring system 2, where the eye movement information base system 9 in the emotion analyzing system 3 compares eye movement information by combining the eye movement information comparing system 10, and can store current eye movement information to facilitate comparison in subsequent eye movement information acquisition, and has a function of increasing comparison information, thereby implementing data learning. The emotion analysis system 3 further comprises an emotion information base system 11 and an emotion information comparison system 12 which is used for performing comprehensive comparison by combining data in the emotion information base system 11 and the eye movement monitoring system 2, wherein the emotion information base system 11 is used for achieving emotion weighting of composite current eye movement information by combining the emotion information comparison system 12, and weight proportion is provided for reference.
As shown in fig. 2, the pupil tracking system 5 includes a track imaging system 501 for recording a pupil movement track, an eye jump frequency monitoring system 502 for monitoring a pupil jump frequency, an eye jump distance monitoring system 503 for monitoring and recording a pupil jump distance in combination with the eye jump frequency monitoring system 502, and a displacement prediction system 504 for predicting a pupil displacement before opening eyes in combination with the blink monitoring system 7 and the eyelid sag monitoring system 8, where the pupil tracking system 5 records eye movement data, and the displacement prediction system 504 can predict and capture an approximate position of a pupil in advance in a blink state, so as to facilitate rapid positioning of the pupil position when an eyelid is opened.
As shown in fig. 3, the blink monitoring system 7 is configured for blink detection, wherein the blink monitoring system 7 comprises a blink rate monitoring system 701 configured for monitoring blink rate, a blink rate monitoring system 702 configured for monitoring blink frequency and a blink time recording system 703 configured for monitoring blink time, wherein the blink rate monitoring system 701 and the blink rate monitoring system 702 are combined to perform blink rate data recording, and detection of which is averaged to be a, blink rates above a are recorded as Aa, and blink rates below a are recorded as Ab. The blink rate and blink frequency are measured over the same time span, and the blink rate measurements are averaged as A, with blink rates above A being recorded as Aa and blink rates below A being recorded as Ab. The blink frequency detection is averaged to obtain a mean value V, and blink frequencies above V are recorded as Va, and blink frequencies below V are recorded as Vb. Through the detection and tracking of the blink speed and the blink frequency, a computable value is given to the thinking activity state, the tensity, the interest degree of the user in the picture, the attention concentration degree and the following performance of the user, so that a quantitative index is given to the blink to provide a quantitative reference value on the emotional change of the user. Meanwhile, the method can also be used as a feature map in later period such as the neural convolution calculation.
While the emotion recognition system for an eye tracker of the present invention has been described in detail with reference to the preferred embodiments thereof, in another aspect, the present invention provides a method for emotion recognition using the emotion recognition system. This is described in detail below in conjunction with the emotion recognition system of FIGS. 1-3.
According to another embodiment of the invention, the emotion recognition method for the eye tracker comprises the following steps:
s1, starting the picture playing system 1 and starting the light supplementing system 4 to supplement light on the surface of the human face, and acquiring eye movement information within a certain time through the eye movement monitoring system 2 within a light ray range meeting the identification requirement; the eye movement information comprises five major information, namely pupil information, blink information, eyelid information, eye jump information, fixation information and the like. In a preferred embodiment, the eye movement information further comprises iris information.
S2, comparing the eye movement information obtained in the S1 with the eye movement information database system 9 and the eye movement information comparison system 10 to obtain basic comparison data; the eye movement information acquired in S1 is represented quantitatively by using the above five pieces of information about the user' S emotion change, and the data is compared with the eye movement reference information stored in the eye movement information library system 9 to form comparison data in the eye movement information comparison system 10 as basic comparison data, thereby visualizing the eye movement information. In this embodiment, the data alignment content in S2 includes pupil trajectory alignment, eye jump frequency and eye jump distance in combination with the pupil tracking system 5, and further includes pupil size alignment in combination with the pupil size monitoring system 6, blink speed alignment and blink frequency alignment in combination with the blink monitoring system 7, and eyelid sag alignment in combination with the eyelid sag monitoring system 8.
And S3, comparing the basic comparison data obtained in the S2 with the emotion information database system 11 and the emotion information comparison system 12, performing weight analysis with different emotion data, and obtaining analysis results with different weights. When the user is happy, the blinking speed becomes fast, the attention is focused, the pupil is bright and the tension of the eyelid is high, and optimistic weighting is obtained, and on the contrary, when the user is depressed, the eyelid is heavy, the eyes are not bright, the attention is distracted, and pessimistic weighting is obtained. By weighting in this manner, emotion is quantized to realize emotion recognition.
The present invention has been described in detail with reference to the specific embodiments, and it will be understood by those skilled in the art that various changes in the embodiments and/or equivalent arrangements of parts of the features of the embodiments may be made without departing from the spirit and scope of the invention, which is defined by the appended claims.