A Novel Face Recognition Approach Using a Multimodal Feature Vector

This paper proposes a face recognition technique using a composite framework consisting of a number of parameters for facial feature extraction and matching. Recognition of a known face often becomes an intricate task due to the presence of lighting variations, viewpoint changes, facial expression variations and presence of occlusions like spectacles or beard. Often, the recognition system may be required to function in the presence of a very small training set; even a single training image in many situations. The feature extraction technique should thus have the potential to extract maximum information from the existing training set and also prove to be resilient in any varying conditions.

Provided by: International Journal of Advances in Engineering & Technology (IJAET) Topic: Security Date Added: Mar 2012 Format: PDF

Find By Topic