KINSHIP VERIFICATION SYSTEM BASED ON MID LEVEL FEATURES

Abstract

Recognizing of human face from image set has recently seen its prosperity because of its effectiveness in dealing with variations
in illumination, expressions, or poses. Unlike most previous kinship verification methods which apply low-level hand-crafted
descriptors such as local binary pattern and Gabor features for face representation. This paper investigates about kinship
verification system based on mid-level features. Here Scale invariant feature transform (SIFT) descriptors have been extracted.
For better performance two features such as blob and corners are detected via Surf and Harris algorithms. Further processed
features are undergone for feature learning algorithm continued with classification. Finally, accuracy rate is to be evaluated for
verification of kinfaces.