Human emotions recognition using adaptive sublayer compensation and various feature extraction mechanisms

Human emotions are an important part of our life. We express our feelings through emotions. Recognition and validation of human emotions have become important for improving the overall human computer interaction. Emotion Recognition is a progressive research area and plays an important role in Computer interaction. For any Emotions Recognition, it is necessary to obtain the image features of face that can be used to detect the emotion class. We implement the adaptive sub-layer compensation (ASLC) based facial emotions recognition method for human emotions recognition with various features extraction mechanism. We modify Marr-Hildreth algorithm by using Adaptive sub-layer compensation and hysteresis analysis to minimize negative effects of Laplacian of Gaussian (LoG), such as image degradation, high response to unwanted details in image, and disconnected edge details from given image. We have studied four different features extraction techniques to identify the emotion. The four feature detection mechanisms implemented are Principal Component Analysis, Local Tetra Pattern, Magnitude Pattern and Wavelet features. The emotion class is identified by using the extracted features and K nearest neighbor algorithm. We identify five different emotion classes i.e. Happy, Sad, Neutral, Surprise and Fear. We have used still image database for the experiments and the results gives identification of input into the five different emotion classes i.e. Happy, Sad, Neutral, Surprise and Fear.