Presentation is loading. Please wait.

Presentation is loading. Please wait.

Analysis of Physiological Responses from Multiple Subjects for Emotion Recognition 2012 IEEE 14th International Conference on e-Health Networking, Applications.

Similar presentations


Presentation on theme: "Analysis of Physiological Responses from Multiple Subjects for Emotion Recognition 2012 IEEE 14th International Conference on e-Health Networking, Applications."— Presentation transcript:

1 Analysis of Physiological Responses from Multiple Subjects for Emotion Recognition 2012 IEEE 14th International Conference on e-Health Networking, Applications and Services(Healthcom) Yuan GU, Kai-Juan WONG, Su-Lim Tan School of Computer Engineering Nanyang Technological University Chairman: Hung-Chi Yang Presenter: Yu-Fang Wu Advisor: Dr. Chun-Ju Hou Date:2014/12/31 1

2 Introduction ◦ Emotions various findings from neuroscience, cognitive science and psychology ◦ Human intelligence development ◦ Social interaction ◦ Perception ◦ Learning ◦ Human’s affective relative to mental disorders ◦ Emotional disorders ◦ Anxiety attacks ◦ Depressions 2

3 Introduction Evaluation Method Name Measurement signal Evaluation Method Emotion Recognition rate Nasoz SC,HR, TemperatureMBP Anger,Sadness Surprise, Fear Amusement Frustration 83.7% Wagner ECG,SC,EMG Respiration SFS & DLS Anger, Joy Sadness Pleasure 92.05% Haag BVP,ECG,SC,EMG Respiration Neural network Valence:89.9% Arousal:96.6% 3

4 Data acquisitions for physiological signals ◦ The ProComp Infiniti from Thought Technology Experimental Setup And Data Collection Channel Sampling Frequency Physiological signal Two channel2048HzBVP, ECG Four lower-speed channel 256HzSC, Rsp, EMGz, EMGc 4

5 Experimental Setup And Data Collection Experimental procedure ◦ Pre-test ◦ Actual test ◦ Post-test 5

6 Methodology Measured features from 6 physiological signals ◦ 50 features (f 1 ~ f 50 ) ◦ HR (f 1-13 ) ◦ BVP (f 14-21 ) ◦ SC (f 22-27 ) ◦ EMGz (f 28-33 ) ◦ EMGc (f 34-39 ) ◦ Rsp (f 40-50 ) 6

7 Results And Discussion Classification method ◦ Linear Discriminant Analysis (LDA) combined with Sequential Floating Forward Search(SFFS) ◦ 10-fold cross validation 7

8 Results And Discussion Result ◦ This research is consistent with Haag's 8 Valencearousal Haag89.9%96.6% this study83.93%85.96%

9 Conclusion This demonstrates that our multi-user system is able to achieve acceptable results when comparable to single-user systems. 9


Download ppt "Analysis of Physiological Responses from Multiple Subjects for Emotion Recognition 2012 IEEE 14th International Conference on e-Health Networking, Applications."

Similar presentations


Ads by Google