Facial expression recognition using modified Viola-John's algorithm and KNN classifier

摘要

In the way of communication, facial expression act as non-verbal communication and play an important role in social interaction by providing some contextual information. Facial expressions also express human's inner emotional state, which is very effective for communication with the actual emotions. In this paper, an algorithm has been proposed to detect the face and facial parts more accurate to the Viola - John's algorithm, and a fast-tracking algorithm for face tracking in real-time scenarios. The fusion of the facial features is used for feature extraction and comparative work on the several classifiers has been presented. In this approach, the images were acquired and seven significant facial parts from the image were cropped, then extract and store the features of several facial expressions. Finally, the expressions in the images were recognized using the classifiers. The algorithm was tested on four kinds of database and achieved accurate performance through the designed system.