Online Unsupervised Feature Learning for Visual Tracking

Feature encoding with respect to an over-complete dic-tionary learned by unsupervised methods, followed by spa-tial pyramid pooling, and linear classification, has exhib-ited powerful strength in various vision applications. Here we propose to use the feature learning pipeline for vi-sual tracking. Tracking is implemented using tracking-by-detection and the resulted framework is very simple yet ef-fective. First, online dictionary learning is used to build a dictionary, which captures the appearance changes of the tracking target as well as the background changes. Given a test image window, we extract local image patches from it and each local patch is encoded with respect to the dictio-nary. The encoded features are then pooled over a spatial pyramid to form an aggregated feature vector. Finally, a simple linear classifier is trained on these features. Our experiments show that the proposed powerful— albeit simple—tracker, outperforms all the state-of-the-art tracking methods that we have tested. Moreover, we eval-uate the performance of different dictionary learning and feature encoding methods in the proposed tracking frame-work, and analyse the impact of each component in the tracking scenario. We also demonstrate the flexibility of feature learning by plugging it into Hare et al.’s tracking method. The outcome is, to our knowledge, the best tracker ever reported, which facilitates the advantages of both fea-ture learning and structured output prediction. 1