Deep Ordinal Hashing with Spatial Attention

Abstract

Hashing has attracted increasing research attentions in recent years due to its high efficiency of computation and storage in image retrieval. Recent works have demonstrated the superiority of simultaneous feature representations and hash functions learning with deep neural networks. However, most existing deep hashing methods directly learn the hash functions by encoding the global semantic information, while ignoring the local spatial information of images. The loss of local spatial structure makes the performance bottleneck of hash functions, therefore limiting its application for accurate similarity retrieval. In this work, we propose a novel Deep Ordinal Hashing (DOH) method, which learns ordinal representations by leveraging the ranking structure of feature space from both local and global views. In particular, to effectively build the ranking structure, we propose to learn the rank correlation space by exploiting the local spatial information from Fully Convolutional Network (FCN) and the global semantic information from the Convolutional Neural Network (CNN) simultaneously. More specifically, an effective spatial attention model is designed to capture the local spatial information by selectively learning well-specified locations closely related to target objects. In such hashing framework,the local spatial and global semantic nature of images are captured in an end-to-end ranking-to-hashing manner. Experimental results conducted on three widely-used datasets demonstrate that the proposed DOH method significantly outperforms the state-of-the-art hashing methods.

References in corpus (10)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Simultaneous Feature Learning and Hash Coding with Deep Neural Networks
• End-to-End Comparative Attention Networks for Person Re-identification
• Bit-Scalable Deep Hashing with Regularized Similarity Learning for Image Retrieval and Person Re-identification
• Supervised Learning of Semantics-Preserving Hash via Deep Convolutional Neural Networks
• Supervised Hashing Using Graph Cuts and Boosted Decision Trees
• The Power of Asymmetry in Binary Hashing
• Deep Region Hashing for Efficient Large-scale Instance Search from Images

Cited by in corpus (4)

• Deep Semantic Multimodal Hashing Network for Scalable Image-Text and Video-Text Retrievals
• HAR-Net: Joint Learning of Hybrid Attention for Single-stage Object Detection
• Asymmetric Deep Semantic Quantization for Image Retrieval
• Extreme Low Resolution Activity Recognition with Confident Spatial-Temporal Attention Transfer