Author : Jeny George

Facial Expressions play an important role in recognizing the human emotions without any verbal communication. Face emotion recognition is one of the main applications of computer vision. The research of emotion recognition includes facial expressions, voice recognition, gesture recognition, text etc. For efficient human-computer interaction, recognizing human emotional state is an important component. The techniques for recognizing facial expression play an important role in monitoring people with mental problems, neuro-developmental disorders, etc. Facial expression recognition systems mainly consists of 3 main parts: face detection, feature extraction and classification. Once the face is detected, the facial feature regions such as eyes, eyebrows, mouth are extracted. Based on the extracted features, expressions are classified. The proposed method is based on information contained in color facial images. The face area is detected from the input image and it is normalized. The purpose of color normalization is to reduce the lighting effect. Features are extracted from the normalized image using log gabor filters. A bank of 24 Log-Gabor filters is used to extract the facial features. Six scales and four orientations are implemented to extract features from face images. These features are then used to detect the facial expressions. The system also detects the facial expressions from blurred images.

Keywords: Facial Expression, Normalization, Feature extraction, Gabor filter

Article published in International Journal of Current Engineering and Technology, Vol.5, No.2 (April-2015)