THE INTERNATIONAL JOURNAL OF SCIENCE & TECHNOLEDGE An Efficient Algorithm for Face and Expression Recognition

Facial expressions convey non-verbal cues. Automatic recognition of facial expressions can be an important component of natural human-machine interfaces. The face is divide into the several regions, and the distribution of the Scale Invariant Feature Transformation (SIFT) features are extracted from them. Facial feature vectors are generated from key point descriptors using Speeded-Up Robust Features. The descriptor can be performed under illumination, noise, expression, and time lapse variations. SIFT detects and uses a much larger number of features from the images, which reduces the contribution of the errors caused by these local variations. When compared with the local Directional Number Pattern, SIFT can increase the efficiency and accuracy in the recognition rate of the face and expression rate.