A robust approach based on local feature extraction for age invariant face recognition

Age variation is a major problem in the area of face recognition under uncontrolled environment such as pose, illumination, expression. Most of the works of this area are based on discriminative methods. The most of the discriminative methods used encoding based descriptors. The encoding based descriptors skip pixels of a few radii width which holds important discriminative information for face recognition under age variation. This paper involves all the pixels of local regions and introduces local descriptors-local difference pattern and local directional gradient relation pattern for extracting texture features for face recognition under age variation. The proposed descriptors are used for robust feature extraction from face images and its parts-periocular region i.e. left and right eye, nose and mouth region. The proposed local difference pattern finds the texture feature difference relation on a local region while local directional gradient relation pattern extracts the feature through finding relation of directional gradient of the local region. Histogram is computed for extracted features using descriptors. Chi-square metric computes the the similarity between probe and gallery images. Experiments have been done on two standard challenging datasets to measure the performance. The proposed approach performed well and presented the better and comparable results i.e. 90.75% recognition accuracy on FGNET and 96.95% recognition accuracy on MORPH.