Sample reconstruction with deep autoencoder for one sample per person face recognition

• Published in: IET computer vision Vol. 11; no. 6; pp. 471 - 478
• Main Authors: Zhang, Yan, Peng, Hua
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 01.09.2017; Wiley
• Subjects: AR database; CDA; class-specific deep autoencoder; CMU database; extended Yale face database B; face recognition; image classification; image coding; image reconstruction; image representation; minimum L2 distance; multisample subject to single-sample subject intraclass variations; one sample per person face recognition; OSPP face recognition; principal component analysis; principle component analysis; regression analysis; Research Article; sample reconstruction; softmax regression; sparse represented-based classifier; CDA; one sample per person face recognition; image classification; AR database; softmax regression; class-specific deep autoencoder; regression analysis; sparse represented-based classifier; minimum L2 distance; image reconstruction; CMU database; extended Yale face database B; principle component analysis; image representation; face recognition; OSPP face recognition; multisample subject to single-sample subject intraclass variations; image coding; sample reconstruction; principal component analysis
• ISSN: 1751-9632; 1751-9640; 1751-9640
• DOI: 10.1049/iet-cvi.2016.0322

One sample per person (OSPP) face recognition is a challenging problem in face recognition community. Lack of samples is the main reason for the failure of most algorithms in OSPP. In this study, the authors propose a new algorithm to generalise intra-class variations of multi-sample subjects to single-sample subjects by deep autoencoder and reconstruct new samples. In the proposed algorithm, a generalised deep autoencoder is first trained with all images in the gallery, then a class-specific deep autoencoder (CDA) is fine-tuned for each single-sample subject with its single sample. Samples of the multi-sample subject, which is most like the single-sample subject, are input to the corresponding CDA to reconstruct new samples. For classification, minimum L2 distance, principle component analysis, sparse represented-based classifier and softmax regression are used. Experiments on the Extended Yale Face Database B, AR database and CMU PIE database are provided to show the validity of the proposed algorithm.