Revisiting Winner Take All (WTA) Hashing for Sparse Datasets
This addresses a problem for researchers and practitioners in computer vision using hashing for large-scale applications, but it is incremental as it improves an existing method.
The paper identified a subtle issue in Winner Take All (WTA) hashing that limits its discriminative power on sparse datasets, and proposed Densified WTA Hashing, which provably fixes the issue and outperforms Vanilla WTA in image classification and retrieval tasks.
WTA (Winner Take All) hashing has been successfully applied in many large scale vision applications. This hashing scheme was tailored to take advantage of the comparative reasoning (or order based information), which showed significant accuracy improvements. In this paper, we identify a subtle issue with WTA, which grows with the sparsity of the datasets. This issue limits the discriminative power of WTA. We then propose a solution for this problem based on the idea of Densification which provably fixes the issue. Our experiments show that Densified WTA Hashing outperforms Vanilla WTA both in image classification and retrieval tasks consistently and significantly.