KR20180122828A - mothod of recognizing emotion by using heart beat and inspiration - Google Patents

Abstract

According to the present invention, a method for recognizing an emotion capable of objectively and quantitatively recognizing an emotion comprises the steps of: detecting electrocardiogram data and breathing data from a subject; generating a breathing rate variability (BRV) spectrum by using the breathing data; generating a heartrate variability (HRV) spectrum from the electrocardiogram data; performing a synchronization analysis on the BRV spectrum and the HRV spectrum by frequency bands; and determining an emotion of the subject according to a value obtained by the synchronization analysis.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a method and system for recognizing emotion using heart and respiration,

The present invention relates to a sensibility recognition method using the connection between heart and breath and a system for applying the same.

Human beings come in contact with people in various situations and live in communication. Communication is to convey the thoughts and meanings of oneself. To communicate effectively, express your own emotions that change depending on the situation, and accurately recognize the emotions of the other person. Therefore, the ability to express and recognize emotion plays a decisive role in the formation of human relationships.

In recent years, ICT technology has attracted attention as an innovative technology of computer and electronic industry technology, which applies anthropomorphism technology to products and services to communicate emotional communication between people, products, people and services beyond human relations. Emotion information communication technology is a technology that recognizes and processes the emotional state of a user who uses a product or service and provides a service suited to the situation. It combines human emotions with products or services to create high value-added It is attracting attention as a technology that can be. In addition, the emotional ICT industry is expected to fuse with various non-ICT industries to create new markets. Sensitivity It is important to accurately recognize emotion in order to provide service according to user's situation through ICT technology.

The method of recognizing emotion can be divided into emotion recognition method according to external reaction and emotion recognition method according to internal reaction. There is a limit to how emotional recognition methods based on external responses are influenced by the cultural and social differences experienced by the perceived subjects. As a research method to overcome these limitations, there is an emotional recognition method according to the physiological signal which is an internal reaction. Sensibility is the internal reaction of the human body caused by the external sensory stimulation, accompanied by a physiological reaction that is an involuntary reaction. Therefore, emotion recognition method using physiological signals has an advantage that it is possible to measure objectively and it is easy to quantify. There are few studies on the relationship between the physiological response parameters and emotional evaluation when emotions are expressed.

The present invention relates to a sensibility recognition method using the connectivity of heart and breath, which can objectively and quantitatively recognize sensibility through connection of heart and respiration under the influence of autonomic nervous system in consideration of the mechanism in the human body, I would like to propose.

The emotion recognition method according to the present invention:

Detecting electrocardiogram data and respiration data from a subject;

Generating a Breathing Rate Variability (BRV) spectrum using the breathing data;

Generating a heart rate variability (HRV) spectrum from the electrocardiogram data;

Performing synchronization analysis on the BRV spectrum and the HRV spectrum on a frequency band basis; And

And determining the emotion of the subject according to the value obtained by the synchronization analysis.

According to a specific embodiment of the present invention, the BRV spectrum can be obtained from IRV (inspiration rate variability) and ERV (Expriation rate vatiability) obtained from the inhalation and exhalation information of the respiration.

According to a specific embodiment of the present invention, the inspiration data having the inspiration time and the expiration time are obtained from the RSP data, and the inspiration data and the expiration data are subjected to an FFT analysis process IRV / ERV spectral data can be extracted from the data.

According to a specific embodiment of the present invention, the frequency band may be divided into VLF (very low frequency, 0.0033-0.04 Hz), LF (low frequency, 0.04-0.15 Hz), and HF (high frequency, 0.15-0.4 Hz) have.

According to a specific embodiment of the present invention, a synchronization analysis by one-way ANOVA is performed, and the sensitivity of the subject is awakened by the difference in synchronization between the LF band and the HF band of the HRV spectrum and the BRV spectrum (Arousal) or relaxation (Relaxation), or Negative or Positive.

According to a specific embodiment of the present invention, the synchronization of the HF band of the HRV spectrum with the BRV spectrum can be analyzed and the emotion of the Arusal or Relaxation evaluated according to the result.

According to a specific embodiment of the present invention, the synchronization of the LF band of the HRV spectrum with the BRV spectrum can be analyzed and the sensitivity of the negative or positive depending on the result can be evaluated.

The emotion recognition apparatus according to the present invention performing the above method comprises:

An electrocardiogram sensor for detecting the electrocardiogram data;

A breathing sensor for detecting the breathing data;

And a computer-based data processing unit for processing data from the electrocardiogram sensor and the breathing sensor.

FIG. 1 is a flow chart for explaining the emotion recognition method according to the present invention, and explains a procedure of extracting HRV and BRV.
2 is a graph showing a process of generating a BRV spectrum according to the present invention.
FIG. 3 shows the result of analyzing the HRV spectrum and the HF band of the BRV spectrum according to the present invention.
FIG. 4 shows a result of analysis of synchronization between the HRV spectrum and the LF band of the BRV spectrum according to the present invention.
FIG. 5 shows a synchronization result pattern in the HF band of Arousal-Relaxation according to the present invention.
FIG. 6 shows a synchronization result pattern in a Negative-Positive LF band according to the present invention.

Hereinafter, embodiments of the emotion recognition method according to the present invention will be described in detail with reference to the accompanying drawings.

The emotion recognition apparatus according to the present invention for performing the data method described below includes: an electrocardiogram sensor for detecting the electrocardiogram data; A breathing sensor for detecting the breathing data; And a computer-based data processing unit for processing data from the electrocardiogram sensor and the breathing sensor. Such a device can be implemented by a general PC in which data processing software as described later is installed and the sensor is connected.

1 is a flowchart of a sensibility recognition method according to the present invention. An experiment method for a subject who verifies the emotion recognition method of the present invention will be described.

<Subject>

Eighteen subjects (10 males, mean age 24.72 +/- 2.47 years) participated in the experiment of the present invention. Subjects participated in the experiment were physically healthy persons without autonomic nervous system history or abnormality.

From 12 hours before participating in the experiment, we restricted caffeine, alcohol, and smoking that could affect the autonomic nervous system.

Except for the purpose of the experiment, the subjects were sufficiently informed about the safety of the measurement of the autonomic nervous system response and the volunteer consent was required to fill out the consent form to participate in the experiment. In addition, a predetermined compensation was paid to increase the participation rate of the experiment.

<Experimental Method>

Subjects participating in the experiment performed listening tasks focused on sound stimuli. The presented stimuli were arousal, relaxation, negative, positive sound stimulation and the experiment was repeated 4 times for 6 minutes each. The sound stimuli presented at the time of repetition did not overlap with the previously presented sound stimuli but randomly selected to eliminate the order effect. The experiment was started, and 3 minutes was set as a section for measuring the bio-signal reference per subject. Then, 3 minutes was required to focus on the sound stimulus presented to the subject, so that the target emotion could be induced.

Electrocardiogram (ECG) and electrocardiogram were attached and measured based on standard limb induction method (Lead I). Respiration (RSP) was measured by wearing a belt-type RSP sensor on the subject's chest.

<Analysis method>

Electrocardiogram and respiration signals were amplified by ECG100C and RSP100C amplifiers (BIOPAC Systems Inc., USA) and digitized with NI-DAQ-Pad9205 (National Instrument Inc., USA) and collected at 500 Hz. The collected electrocardiogram and respiratory signals were signaled through LabVIEW 2010 (National Instrument Inc., USA) software.

The measured electrocardiogram data were subjected to R-peak (R-peak) detection by R-peak detection using a peak detection algorithm. RRI data were extracted from HRV spectral data through FFT (fast fourier transform) analysis. The extracted HRV spectra are divided into VLF (very low frequency, 0.0033 - 0.04 Hz), LF (low frequency, 0.04 - 0.15 Hz) and HF (high frequency, 0.4 Hz), respectively.

, inspiratory time)으로 저장하고, 포지티브 피크(Positive peak)와 이에 이어지는 네가티브 피크(Negative peak) 사이의 간격은 날숨 간격(

, expiratory time)으로 저장하였다.

와

데이터는 각각 FFT 분석을 통해 IRV (inspiration rate variability) 와 ERV (expiration rate variability) 스펙트럼 데이터를 추출하였다.Measured respiration (RSP) data detected positive and negative peaks related to inspiration and exhalation through a peak detection algorithm. The interval between the detected negative peak and the following positive peak is defined as the interval between the in-

, inspiratory time), and the interval between the positive peak and the following negative peak is stored as the exhalation interval (

, and expiratory time.

Wow

Data were extracted from the inspiration rate variability (IRV) and the expiration rate variability (ERV) spectral data through FFT analysis, respectively.

2 is a graph showing a process of extracting a BRV spectrum according to the present invention. In FIG. 2, the first and second graphs are the IRV spectrum and the ERV spectrum showing the change in the power spectral density (PSD), and the last graph shows the BRV spectrum obtained using the IRV spectrum and the ERV spectrum.

In the heart and respiration, VLF, LF band and inspiration are utilized as sympathetic activation index, and HF band and exhalation are utilized as parasympathetic activation index.

Therefore, a new BRV (breathing rate variability) spectrum was generated by connecting the VLF, LF band of the IRV spectrum and the HF band data of the ERV spectrum. The band of extracted BRV spectrum was divided into the same as the HRV spectrum band of ECG, and the ratio of VLF, LF and HF was obtained.

와

데이터를 시계열 데이터로 변환하여 FFT분석을 진행하기에 동일한 대역에서 신호처리 및 분석이 가능하다.Heart rate and respiration are different between the heart rate and respiratory rate, but RRI,

Wow

Data is transformed into time-series data and the FFT analysis is performed, so that signal processing and analysis can be performed in the same band.

When sensory changes are caused by external sensory stimuli, they are accompanied by an innate response, an involuntary physiological response. At this time, the heart and respiration maintain the homeostasis through the organic action of the sympathetic nerve and the parasympathetic nerve. Therefore, the connectivity (synchronization) between heart and respiration was calculated by the difference in the power amount (PSD) of the LF and HF frequency bands of HRV and BRV spectra.

<Statistical Analysis>

(partial eta-squared)와 d (cohen's d)을 통해 계산하였다. 효과 크기의

값은 일반적으로 0.02, 0.13, 0.26에 따라 낮은 수준, 중간 수준, 높은 수준을 의미한다. 효과 크기의 d 값은 0.20, 0.50, 0.80에 따라 낮은 수준, 중간 수준, 높은 수준을 의미한다. 모든 통계 분석은 SPSS 21 (IBM, USA)을 사용해 분석하였다.As a result of regularity test, the difference in synchronization between LF band and HF band of electrocardiogram and respiration spectrum according to emotion satisfied regularity. Therefore, statistical significance was analyzed through one-way ANOVA and post-test was analyzed by Tukey test. To solve the multiple comparison problem, Bonferroni's correction was applied to reset the statistical significance level to 0.025 (α = 0.05 / 2) (Dunnett, 1955). The size of the effect used for substantive significance verification

(partial eta-squared) and d (cohen's d). Effect size

The values generally refer to low, medium and high levels according to 0.02, 0.13, and 0.26. The d value of effect size means 0.20, 0.50, 0.80, low, medium, high level. All statistical analyzes were performed using SPSS 21 (IBM, USA).

<Statistical Analysis Results>

= 0.1372, LF 대역의 동기화: F(3, 68) = 5.075, p = 0.003,

= 0.1823). In the experiment of the present invention, a statistically significant difference was found in the difference of synchronization between the LF band and the HF band of HRV spectrum and BRV spectrum according to sensitivity (Arousal, Relaxation, Negative, Positive) through one-way ANOVA (HF band synchronization: F (3, 68) = 3.605, p = 0.018,

= 0.1372, synchronization of the LF band: F (3, 68) = 5.075, p = 0.003,

= 0.1823).

The post test was conducted by Tukey test to confirm the significance between arousal, relaxation, negative, and positive emotion. The results of the synchronization analysis of the HRF spectrum between the sensibility and the HF band of the BRV spectrum are as shown in Fig.

As shown in FIG. 3, the arousal showed a statistically significant difference from the relaxation, but no statistically significant difference was found between the negative and positive [Arousal (Relaxation: p = 0.008, d = 1.4274, Negative: p = 0.389, d = 0.4174, Positive: p = 0.365, d = 0.4395). Relaxation also did not reveal negative, positive, or statistically significant differences [Relaxation (Negative: p = 0.334, d = 0.4696, Positive: p = 0.358, d = 0.4462). Negative also did not identify a statistically significant difference from other emotions [Negative (Positive: p = 1, d = 0)].

The analysis result of the synchronization between the HRV spectrum of the sensibility and the LF band of the BRV spectrum is as shown in FIG.

Arousal (Anger) was not statistically significant different from other emotions (Arousal (Relaxation: p = 0.593, d = 0.2571, Negative: p = 0.103, d = 0.8152, Positive: p = 0.525, d = 0.3063) ]. Relaxation also did not show statistically significant difference from other emotions (Relaxation (Negative: p = 0.712, d = 0.1772, Positive: p = 0.051, d = 0.9945). Negative, however, was positively and statistically significant [Negative (Positive: p = 0.002, d = 1.7377)].

<Analysis result of emotional pattern>

FIG. 5 shows a synchronization result pattern in the HF band of Arousal-Relaxation.

Synchronization of HRV spectrum and HF band of BRV spectrum showed statistically significant results between Arousal and Relaxation emotions. The synchronization result was normalized because subjects could have individual differences in the baseline. In the HF band synchronization, 14 subjects except 4 subjects showed a high Arousal and a low Relaxation based on the synchronization value of 0.5.

Figure 6 shows a pattern of synchronization results in the Negative-Positive HF band.

As shown in FIG. 6, the synchronization between the HRV spectrum and the LF band of the BRV spectrum showed statistically significant results between the negative and positive emotions. In the LF band synchronization, 15 subjects (excluding 3 subjects) showed a low negative value and a positive positive pattern at 0.5.

As described above, the present invention provides an emotional recognition method capable of objectively and quantitatively recognizing emotions through the connection between the heart and the respiration affected by the autonomic nervous system in consideration of the mechanism in the human body.

Recently, emotional ICT technology has been attracting attention. Emotion information communication technology recognizes and processes emotional state of users who use a product or service, It is important to recognize emotion accurately in order to provide accurate service.

Therefore, according to the present invention, the sensibility can be recognized objectively and quantitatively through the connection of the heart and the respiration, which is expected to be utilized in various industrial fields.

While the invention has been shown and described with reference to certain preferred embodiments thereof, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (8)

Detecting electrocardiogram data and respiration data from a subject;
Generating a Breathing Rate Variability (BRV) spectrum using the breathing data;
Generating a heart rate variability (HRV) spectrum from the electrocardiogram data;
Performing synchronization analysis on the BRV spectrum and the HRV spectrum on a frequency band basis; And
And determining the emotion of the subject according to the value obtained by the synchronization analysis. The method according to claim 1,
Wherein the BRV spectrum is obtained from an inspiratory rate variability (IRV) and an expiratory rate vatiability (ERV) obtained from the inhalation and exhalation information of the respiration. The method according to claim 1,
And extracting IRV / ERV spectral data through the FFT analysis of the inspiration data and the expiration data by obtaining expiratory data having inspiration time and expiration time from the respiration data A Sensory Recognition Method Using the Connectivity of Heart and Respiration as Features. The method according to claim 1,
The frequency band is divided into VLF (very low frequency, 0.0033-0.04 Hz), LF (low frequency, 0.04-0.15 Hz) and HF (high frequency, 0.15-0.4 Hz) Sensibility recognition method used. The method according to claim 1,
A one-way ANOVA is performed to perform a synchronization analysis, and thereby, the sensitivity of the subject is arousal or relaxation, and the sensitivity of the subject is affected by the difference in synchronization between the HRF spectrum and the BRF spectrum in the LF band and the HF band, Or negative or positive of the heart, and determining the sensation using the connectivity of the heart and the respiration. 6. The method of claim 5,
Analyzing the synchronization of the HRV spectrum and the HF band of the BRV spectrum and evaluating the emotion of arness or relaxation according to the result of the analysis. 6. The method of claim 5,
Analyzing the synchronization of the HRV spectrum and the LF band of the BRV spectrum and evaluating the sensibility of a negative or positive according to the result of the analysis. 8. An emotion recognition apparatus using heart and respiration connectivity to perform the method according to any one of claims 1 to 7,
An electrocardiogram sensor for detecting the electrocardiogram data;
A breathing sensor for detecting the breathing data;
And a computer-based data processing unit for processing data from the electrocardiogram sensor and the breathing sensor.

Images