| Emotion plays a significant role in human decision and perception making. For long a time, Research on emotion intelligence has been done in the fields of psychology and cognitive science. Along with the development of artificial intelligence these years, the combination of emotion intelligence and computer technology brings the novel research area named affective computing. To a large extent, this combination will promote development of the computer technology. Computer automatic recognition of human emotion is the first step toward affective computing. During the facial express, speech, posture and physiological emotion recognition, physiological signals, which are viewed as the most reality emotion express of human emotion, carry plenty of emotion information. How to automatically recognize emotional state from physiological signals is the subject of attention by researchers form different fields. Recent studies on emotion recognition based on physiological signals features, still have some drawbacks. For example, features which can be widely used, like the features used in facial recognition, have not been found. Furthermore, the recognition rate is not high enough to be widely used in practice. Therefore, research on physiological signal's emotion recognition still has larger developing spatial. Especially, select of emotion's features is a developing trend in computational intelligence fields.Tabu search (TS) algorithm is a meta-heuristic algorithm. TS can avoid circuit searching by using the flexible memory mechanism and respective tabu criteria. Also according to aspiration criteria, TS can assoil some good solution status which have been tabued, in doing so it can ensure the diversification search and obtain the globe optimum. Recent researches show that tabu search has the equivalent (even better) capability to Genetic algorithms and simulated annealing. TS has successful applications in combinational optimization method, production scheduling and feature selection algorithms. However, there are very few researches on using tabu search as feature selection method for emotion recognition.Aiming at limitation of features selection of traditional physiological signals, this paper selects emotion's features for recognize emotion using tabu search., by focusing in the four emotional states: happiness, anger, sadness, and pleasure state which can be always found in daily life. The main contents are listed as follows:(1) According to the characteristic of physiological signals itself, 4 physiological signals including Electrocardiogram (ECG), Skin Conductivity (SC), Electromyography (EMG) and Respiration (RSP) were manipulated by difference, filtering, frequently filed transform operation and so on. The 193 features had been found.(2) Combining tabu search and neighbor method, correct rate of classification is viewed as criterion function of tabu search algorithm, which is used for emotion recognition of joy, anger, sadness and pleasure. It is considered that feature selection of physiological and emotion classification operate at the same time. It can be seen that more physiological signals can achive higher classification affect. Emotion recognition from single signals and 4 signals both obtain better recognition rates. Emotion recognition correct rate from single signals is up to 50%. While general emotion recognition rate of 4 signals arrives more that 80%. In the experiment of single emotion recognition, correct rate of angle often get to 100%.(3) Features of physiological signals are selected using tabu search. After many times experiments, it can reach feature subsets of best recognition rate. For example, RSP is the most important signal for 4 emotions in the paper. When it gains the best recognition rate, rspPulse-median is selected.From the results of the research given in this paper, we can see that tabu search has the high recognition rates in emotion recognition research. In addition, tabu search can find the robust feature subsets.
|
PDF Full Text Request