Deep unsupervised learning of visual similarities. (June 2018)

• Record Type: Journal Article
• Authors: Sanakoyeu, Artsiom; Bautista, Miguel A.; Ommer, Björn
• Is Part Of: Pattern recognition. Volume 78(2018:Jun.)
• Page Start: 331
• Page End: 343
• Publication Date: 2018-06
• Subjects: Visual similarity learning -- Deep learning -- Self-supervised learning -- Human pose analysis -- Object retrieval; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2018.01.036
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: Unsupervised visual similarity learning is framed as a surrogate classification task. Use weak estimates of local similarities to group samples into compact cliques. Train a ConvNet to learn visual similarities by learning to categorize cliques. Optimization problem to sample training minibatches without conflicting relations. Competitive performance on detailed posture analysis and object classification. Abstract: Exemplar learning of visual similarities in an unsupervised manner is a problem of paramount importance to computer vision. In this context, however, the recent breakthrough in deep learning could not yet unfold its full potential. With only a single positive sample, a great imbalance between one positive and many negatives, and unreliable relationships between most samples, training of Convolutional Neural networks is impaired. In this paper we use weak estimates of local similarities and propose a single optimization problem to extract batches of samples with mutually consistent relations. Conflicting relations are distributed over different batches and similar samples are grouped into compact groups. Learning visual similarities is then framed as a sequence of categorization tasks. The CNN then consolidates transitivity relations within and between groups and learns a single representation for all samples without the need for labels. The proposed unsupervised approach has shown competitive performance on detailed posture analysis and object classification.