High Impact Factor : 4.396 icon | Submit Manuscript Online icon | UGC Approved icon

Robust Technique of Age and Expression Invariant Feature Extraction

Author(s):

Prof. Md Yousuf Ahmed , KBN College of Engineering, Kalaburagi ; Prof. Feroza D Mirajkar , KBN College of Engineering, Kalaburagi

Keywords:

Face recognition, age invariance, local feature representation, SIFT, MLBP, Multi-Feature Discriminant Analysis, Discriminative Model

Abstract

The development of a robust technique to extract the facial features that is invariant to extrinsic (head, pose, lightning conditions) and intrinsic (aging, facial expression) source of variability is one of the greatest challenges of current research in this field. Designing such a robust technique can overcome the problems of missing child, multiple enrollments where subjects purposely try to hide their identity. In this paper, a robust technique is developed which extracts the features invariant to aging and expression variation. The two local descriptors SIFT and MLBP are used to extract the local features of the image. The random sampling techniques, random subspace and bagging are used to construct the multiple classifiers of lower dimensionality. The final decision is made using the simple score sum based rule.

Other Details

Paper ID: NCACCP032
Published in: Conference 8 : NCACC-2016
Publication Date: 01/10/2016
Page(s): 126-133

Article Preview




Download Article