Title :
Face Verification Using the LARK Representation
Author :
Seo, Hae Jong ; Milanfar, Peyman
Author_Institution :
Univ. of California, Santa Cruz, CA, USA
Abstract :
We present a novel face representation based on locally adaptive regression kernel (LARK) descriptors. Our LARK descriptor measures a self-similarity based on “signal-induced distance” between a center pixel and surrounding pixels in a local neighborhood. By applying principal component analysis (PCA) and a logistic function to LARK consecutively, we develop a new binary-like face representation which achieves state-of-the-art face verification performance on the challenging benchmark “Labeled Faces in the Wild” (LFW) dataset. In the case where training data are available, we employ one-shot similarity (OSS) based on linear discriminant analysis (LDA). The proposed approach achieves state-of-the-art performance on both the unsupervised setting and the image restrictive training setting (72.23% and 78.90% verification rates), respectively, as a single descriptor representation, with no preprocessing step. As opposed to combined 30 distances which achieve 85.13%, we achieve comparable performance (85.1%) with only 14 distances while significantly reducing computational complexity.
Keywords :
computational complexity; face recognition; principal component analysis; regression analysis; LARK representation; binary like face representation; computational complexity; face verification; image restrictive training setting; labeled faces in the wild dataset; linear discriminant analysis; locally adaptive regression kernel descriptors; logistic function; one shot similarity; principal component analysis; signal induced distance; Face detection; Face recognition; Linear discriminant analysis; Principal component analysis; Training; Training data; Face verification; labeled faces in the wild (LFW); locally adaptive regression kernels (LARKs); matrix cosine similarity; one-shot similarity (OSS);
Journal_Title :
Information Forensics and Security, IEEE Transactions on
DOI :
10.1109/TIFS.2011.2159205