نویسندگان
دانشگاه صنعتی سهند - دانشکده مهندسی برق - آزمایشگاه علوم اعصاب محاسباتی
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
Emotion recognition is one of the challenging issues among researchers, particularly in the field of affective computing. Although some success has been achieved in this area, but still no algorithm has been suggested that can recognize human emotions perfectly. Lack of attention to individual differences in response to an emotional stimulus may be one of the reasons. Among them, the most important factor is gender differences in response to emotions. The present study investigated the gender differences through electrocardiogram (ECG), galvanic skin response (GSR), finger pulse signals by considering comprehensive signal processing approaches. The autonomic signals of 47 students (mean age: women 21.9 ± 1.7, and men 21.1 ± 1.48 years) were recorded while watching happy, sadness, fear, and relaxed images. A wide range of features has been extracted from the three signals and differences between autonomic responses of men and women were evaluated by means of Wilcoxon statistical test. The results showed that there are significant differences in physiological responses of men and women while watching affective pictures, where the capability of heart responses is much higher than the GSR. In addition, the importance and effective role of nonlinear indices in the analysis of bio-signals of two genders in response to affective pictures has been proved. The most significant difference in the features extracted from two groups was related to the sadness. The results of this study emphasize on the gender role as an important factor in the problem of emotion recognition and suggest that individual differences should be considered in designing an emotion recognition system.
کلیدواژهها [English]
)