Emotion recognition based on vertical cross correlation sequence of facial expression images
M Imran Rashid, Md. Robiul Hasan, Nilofa Yeasmin, Celia Shahnaz, Shaikh Anowarul Fattah, Wei‐Ping Zhu, Mudassar Ahmed · 2017
Command extraction from human beings becomes easier for a machine if it can analyze the non verbal ways of communication such as emotions. This paper focuses on improving the efficiency of extracting emotion from human facial expression images. The features that were extracted in this experiment were obtained from JAFFE (Japanese Female Facial Expression) database which includes 213 images of different models who posed for 7 classes of expressions. Their images were refined through some steps and prepared for the classification process to recognize the emotions. We used viola-jones algorithm to find the ROI (regions of interest), where human emotions can be observed and detected noticeably and then the ROI parts have been segmented separately after some pre-processing. With these pre-processed segmented parts, we have performed cross correlation to build our feature vector which gives us the information about variation. The k-NN classifier has been used to detect emotion from test images.