An improved method for emotion detection using weighted Gabor filter and radial basis Function kernel

Human emotion detection plays an important role in the human-computer interaction. In this paper, the emotions are detected on segmented image using low dimension weighted Gabor filter bank. The segmentation reduces the space domain and only those facial features are focused that reflects expressions accurately. The classification of selected features values classifies through a RBF network. The experimental results show that by the selection of optimal features, the computational complexity reduces significantly.