Deep Learning to Ternary Hash Codes by Continuation

TL;DR

This paper introduces a novel deep learning approach that jointly learns features and ternary hash codes using a continuation method, resulting in improved image retrieval accuracy over traditional binary codes.

Contribution

It proposes a new joint learning framework with a smoothed function and continuation method to optimize ternary hash codes directly from deep features.

Findings

Ternary codes outperform binary codes in image retrieval accuracy.

The continuation method effectively reduces quantization errors.

The approach achieves higher retrieval precision in experiments.

Abstract

Recently, it has been observed that {0,1,-1}-ternary codes which are simply generated from deep features by hard thresholding, tend to outperform {-1,1}-binary codes in image retrieval. To obtain better ternary codes, we for the first time propose to jointly learn the features with the codes by appending a smoothed function to the networks. During training, the function could evolve into a non-smoothed ternary function by a continuation method. The method circumvents the difficulty of directly training discrete functions and reduces the quantization errors of ternary codes. Experiments show that the generated codes indeed could achieve higher retrieval accuracy.