Face recognition using hybrid classifiers

摘要

We address the problem of surveillance and contents-based image retrieval (CBIR) for large image databases consisting of face images. The corresponding face recognition tasks considered herein include (i) surveying a gallery of images for the presence of specific probes. (ii) CBIR, and (iii) CBIR subject to correct ID (“match”) displaying specific facial landmarks such as wearing glasses. We developed robust matching (“classification”) and retrieval schemes based on hybrid classifiers and showed their feasibility using the FERET database. The hybrid classifier architecture consists of an ensemble of connectionist networks—radial basis functions (RBF)—and inductive decision trees (DT). The specific characteristics of our hybrid architecture include (a) query by consensus as provided by ensembles of networks for coping with the inherent variability of the image formation and data acquisition process, (b) categorical classifications using decision trees, (c) flexible and adaptive thresholds as opposed to ad hoc and hard thresholds, and (d) interpretability of the way classification and retrieval are eventually achieved. Experimental results, proving the feasibility of our approach, yield (i) 96% accuracy, using cross validation, for surveillance on a database consisting of 904 images corresponding to 350 subjects (of whom 102 are duplicates), (ii) 97% accuracy for CBIR tasks, such as “find all subjects wearing glasses”, on a database of 1084 images (including noisy versions) of 350 subjects (of whom 102 are duplicates), and (iii) 93% accuracy, using cross validation, for CBIR subject to correct ID match tasks, such as “find Joe Smith with/without glasses”, on a database of 200 images.