Facial expression recognition of multiple stylised characters using deep convolutional neural network.

Human faces manifest the treasury of their abilities including emotions, character, state of mind and many more. Apart from the things that are spoken, human faces conveys plenty of information in the form of facial expressions. Recognition of facial expressions has become significant in the discipline of human-computer interaction to attain the emotional state of human beings. This paper proposes a facial expression identification method (FEIM) for the recognition of six basic facial expressions (anger, sad, fear, happy, surprise and disgust) plus one neutral emotion. The features are extracted by implementing an integrated Gabor and local binary pattern (LBP) feature extraction method and the concept of principal component analysis (PCA) is executed for feature selection. A deep neural network is trained for the facial expression research group database (FERG-DB) dataset to classify the facial expression images into seven emotion expression classes (anger, fear, disgust, happy, neutral, sad, and surprise). The effectiveness of the proposed system is manifested by comparing the recognition rate results with state-of-the-art-techniques. The overall results in terms of precision, recall and f-score also favours the efficacy of proposed method.