End-to-end Binary Representation Learning via Direct Binary Embedding

TL;DR

This paper introduces Direct Binary Embedding (DBE), an end-to-end deep learning approach that directly learns binary representations from images without quantization errors, improving performance on recognition and retrieval tasks.

Contribution

The paper proposes a novel end-to-end binary embedding method with a special DBE layer and a joint loss for multilabel images, outperforming existing hashing techniques.

Findings

DBE outperforms state-of-the-art methods in image retrieval.

The joint loss effectively handles multilabel image data.

DBE reduces quantization errors in binary representation learning.

Abstract

Learning binary representation is essential to large-scale computer vision tasks. Most existing algorithms require a separate quantization constraint to learn effective hashing functions. In this work, we present Direct Binary Embedding (DBE), a simple yet very effective algorithm to learn binary representation in an end-to-end fashion. By appending an ingeniously designed DBE layer to the deep convolutional neural network (DCNN), DBE learns binary code directly from the continuous DBE layer activation without quantization error. By employing the deep residual network (ResNet) as DCNN component, DBE captures rich semantics from images. Furthermore, in the effort of handling multilabel images, we design a joint cross entropy loss that includes both softmax cross entropy and weighted binary cross entropy in consideration of the correlation and independence of labels, respectively. Extensive experiments demonstrate the significant superiority of DBE over state-of-the-art methods on tasks of natural object recognition, image retrieval and image annotation.