DeepHash: Getting Regularization, Depth and Fine-Tuning Right

TL;DR

DeepHash is a deep network-based hashing scheme that effectively compresses high-dimensional image descriptors into 64-1024 bit binary hashes, achieving near uncompressed retrieval performance with significant compression.

Contribution

The paper introduces DeepHash, a novel deep learning approach for compact binary hashing that emphasizes regularization, depth, and fine-tuning tailored for extremely low bitrates.

Findings

Outperforms state-of-the-art hashing methods by up to 20%.

Achieves retrieval performance close to uncompressed features with 256-bit hashes.

Provides 512 times compression of image descriptors.

Abstract

This work focuses on representing very high-dimensional global image descriptors using very compact 64-1024 bit binary hashes for instance retrieval. We propose DeepHash: a hashing scheme based on deep networks. Key to making DeepHash work at extremely low bitrates are three important considerations -- regularization, depth and fine-tuning -- each requiring solutions specific to the hashing problem. In-depth evaluation shows that our scheme consistently outperforms state-of-the-art methods across all data sets for both Fisher Vectors and Deep Convolutional Neural Network features, by up to 20 percent over other schemes. The retrieval performance with 256-bit hashes is close to that of the uncompressed floating point features -- a remarkable 512 times compression.