Analysis of Human Emotions using Computational Geometry for Nose and Lips Facial Features
Understanding, recognizing facial expressions and extracting meaningful emotions from face is a very challenging task in today’s Era. These expressions can be represented through images/videos. In general image/video data is viewed as big data, processing and analysing this type of data manually is highly tedious. But this process is important as the insights of this data can be proved to be beneficial in terms of security and predictions etc. Thus, Facial emotion analysis and prediction is a challenging research area. In this paper, we present an approach to build the emotion prediction model through facial expression analysis. This approach considers two features in face particularly NOSE and LIPS and captures the geometric structure of them based on various emotions. The approach considers basic six emotions such as Sad, Happy, Anger, Surprise, Fear and Disgust. The process includes pre-processing steps such as face identification from image segmentation, feature extraction. The approach identifies feature vectors of NOSE and LIPS and considers these as landmarks for emotion identification. Later, distance measure is calculated among these feature vectors. The approach is initially processed for Neutral face then extended to the faces containing various emotions. Later, emotion classification is done based on SVM classifier. The approximate success rate for this proposed approach of analyzing the emotions results in an average of 76%.
Keywords - Face; Emotions; Features; Distance Measure; Prediction; Landmarks