DeepQ Residue Analysis of Computer Vision Dataset using Support Vector Machine

Abstract

A vision-based human computer interface is used to automatically recognize human mood. Image processing techniques used include a web camera for eye detection. Appearance tracking method (ABT) is measured face identification and K means Nearest Neighbor (K-NN) is used for eye detection. DWT - Discrete Wavelet Transform and DCT - Discrete Cosine Transform are suitable to extract features of eye and SVM is used to classify eye expressions. Classification of eye expressions includes anger, fear, happiness, disgust, neutral and sad. Experimental results confirm that the proposed method recognized facial expressions with higher accuracy.