Emotion feature fusion of images based on electroencephalogram analysis
Xin Luo, Qiao Pan, Guangwei Xu, Guowen Wu · 2010
A method of emotion feature fusion of images based on EEG analysis is proposed in this paper. At first, select the appropriate algorithm to extract color and shape feature. Then, the EEG analysis is proposed, analyses are made on emotion feature of image by way of Emotion Spectrum Analyzing. The paper uses cloth image as date, proposes the Canonical Correlation Analysis (CCA) method which base on multi-features including the visual feature (color and shape) and the emotional feature. The data set composed of 100 clothing images is experimented with, and the experiment results have been analyzed, some of which are satisfied, and that this method is efficient and robust.