How should we evaluate supervised hashing?

TL;DR

This paper critiques current evaluation methods for supervised hashing, proposes improved protocols based on retrieval and transfer learning, and provides baseline comparisons to better assess hashing performance.

Contribution

It identifies flaws in existing evaluation protocols and introduces two new evaluation methods for supervised hashing, along with baseline comparisons.

Findings

Existing protocols can be bypassed by trivial solutions.

Proposed protocols better reflect real-world performance.

Baseline methods establish performance bounds.

Abstract

Hashing produces compact representations for documents, to perform tasks like classification or retrieval based on these short codes. When hashing is supervised, the codes are trained using labels on the training data. This paper first shows that the evaluation protocols used in the literature for supervised hashing are not satisfactory: we show that a trivial solution that encodes the output of a classifier significantly outperforms existing supervised or semi-supervised methods, while using much shorter codes. We then propose two alternative protocols for supervised hashing: one based on retrieval on a disjoint set of classes, and another based on transfer learning to new classes. We provide two baseline methods for image-related tasks to assess the performance of (semi-)supervised hashing: without coding and with unsupervised codes. These baselines give a lower- and upper-bound on the performance of a supervised hashing scheme.