Multi-scale invariant abstracted under varying illumination
Bin Xu, Taiping Zhang, Zhaowei Shang · 2012
Making recognition more reliable under uncontrolled lighting conditions is one of the most important challenges for face recognition. In this correspondence, multi-scale illumination invariant is derived from the image gradient domain (MGI) which can discover underlying inherent structure while keeping the details at most. The resulting method provides state-of-the-art performance on two data sets that are widely used for testing recognition under difficult illumination conditions: Extended Yale-B and PIE. Recognition rates of 99.11% achieved on PIE database of 68 subjects, 99.38% achieved on Yale B of ten subjects which outperforms most existing approaches.