Simultaneous Feature Learning and Hash Coding with Deep Neural Networks

Abstract

Similarity-preserving hashing is a widely-used method for nearest neighbour search in large-scale image retrieval tasks. For most existing hashing methods, an image is first encoded as a vector of hand-engineering visual features, followed by another separate projection or quantization step that generates binary codes. However, such visual feature vectors may not be optimally compatible with the coding process, thus producing sub-optimal hashing codes. In this paper, we propose a deep architecture for supervised hashing, in which images are mapped into binary codes via carefully designed deep neural networks. The pipeline of the proposed deep architecture consists of three building blocks: 1) a sub-network with a stack of convolution layers to produce the effective intermediate image features; 2) a divide-and-encode module to divide the intermediate image features into multiple branches, each encoded into one hash bit; and 3) a triplet ranking loss designed to characterize that one image is more similar to the second image than to the third one. Extensive evaluations on several benchmark image datasets show that the proposed simultaneous feature learning and hash coding pipeline brings substantial improvements over other state-of-the-art supervised or unsupervised hashing methods.

This paper has been accepted to IEEE International Conference on Pattern Recognition and Computer Vision (CVPR), 2015

References in corpus (4)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Going Deeper with Convolutions
• Learning Hash Functions Using Column Generation

Cited by in corpus (144)

• Deep Recurrent Quantization for Generating Sequential Binary Codes
• Beyond Product Quantization: Deep Progressive Quantization for Image Retrieval
• A Decade Survey of Content Based Image Retrieval using Deep Learning
• Supervised Learning of Semantics-Preserving Hash via Deep Convolutional Neural Networks
• Learning Discrete Representations via Information Maximizing Self-Augmented Training
• Deep Learning Techniques for Future Intelligent Cross-Media Retrieval
• SSDH: Semi-supervised Deep Hashing for Large Scale Image Retrieval
• Deep Supervised Discrete Hashing
• Recent Advance in Content-based Image Retrieval: A Literature Survey
• Instance-Aware Hashing for Multi-Label Image Retrieval
• Deep Semantic Multimodal Hashing Network for Scalable Image-Text and Video-Text Retrievals
• Visual Search at eBay
• Feature Learning based Deep Supervised Hashing with Pairwise Labels
• Deep Discrete Supervised Hashing
• Deep Ordinal Hashing with Spatial Attention
• Deep Hashing for Secure Multimodal Biometrics
• Part-based Deep Hashing for Large-scale Person Re-identification
• Structured Deep Hashing with Convolutional Neural Networks for Fast Person Re-identification
• HashNet: Deep Learning to Hash by Continuation
• Comparative Deep Learning of Hybrid Representations for Image Recommendations
• Deep Triplet Hashing Network for Case-based Medical Image Retrieval
• Deep Learning based Large Scale Visual Recommendation and Search for E-Commerce
• HS-GCN: Hamming Spatial Graph Convolutional Networks for Recommendation
• Video2GIF: Automatic Generation of Animated GIFs from Video
• Unsupervised Deep Cross-modality Spectral Hashing
• Tensor Normalization and Full Distribution Training
• Where Is My Training Bottleneck? Hidden Trade-Offs in Deep Learning Preprocessing Pipelines
• Learning to Hash for Indexing Big Data - A Survey
• Unsupervised Deep Hashing for Large-scale Visual Search
• Fine-grained Categorization and Dataset Bootstrapping using Deep Metric Learning with Humans in the Loop
• Simultaneous Feature Aggregating and Hashing for Compact Binary Code Learning
• Now Playing: Continuous low-power music recognition
• Deep Multi-View Enhancement Hashing for Image Retrieval
• Transductive Zero-Shot Hashing for Multilabel Image Retrieval
• Deep Heterogeneous Hashing for Face Video Retrieval
• Central Similarity Quantization for Efficient Image and Video Retrieval
• From handcrafted to deep local features
• Rank-Consistency Deep Hashing for Scalable Multi-Label Image Search
• HashGAN:Attention-aware Deep Adversarial Hashing for Cross Modal Retrieval
• Correlation Hashing Network for Efficient Cross-Modal Retrieval
• Deep Attention-guided Hashing
• Transitive Hashing Network for Heterogeneous Multimedia Retrieval
• Intermediate Deep Feature Compression: the Next Battlefield of Intelligent Sensing
• Embedding Label Structures for Fine-Grained Feature Representation
• One Loss for All: Deep Hashing with a Single Cosine Similarity based Learning Objective
• Indexing of CNN Features for Large Scale Image Search
• A Revisit on Deep Hashings for Large-scale Content Based Image Retrieval
• Learning to Hash with Graph Neural Networks for Recommender Systems
• IEGAN: Multi-purpose Perceptual Quality Image Enhancement Using Generative Adversarial Network
• Texture Synthesis Guided Deep Hashing for Texture Image Retrieval
• Semantic Cluster Unary Loss for Efficient Deep Hashing
• Deep Hashing with Category Mask for Fast Video Retrieval
• Learning 3D Segment Descriptors for Place Recognition
• Simultaneous Feature Aggregating and Hashing for Large-scale Image Search
• Learning in High-Dimensional Multimedia Data: The State of the Art
• Towards Efficient Unconstrained Palmprint Recognition via Deep Distillation Hashing
• Deep Reinforcement Learning for Image Hashing
• MIHash: Online Hashing with Mutual Information
• Deep Binaries: Encoding Semantic-Rich Cues for Efficient Textual-Visual Cross Retrieval
• Simultaneous Region Localization and Hash Coding for Fine-grained Image Retrieval
• Fast Training of Triplet-based Deep Binary Embedding Networks
• End-to-End Supervised Product Quantization for Image Search and Retrieval
• Tiny Descriptors for Image Retrieval with Unsupervised Triplet Hashing
• Deep Hashing: A Joint Approach for Image Signature Learning
• Self-Supervised Bernoulli Autoencoders for Semi-Supervised Hashing
• Asymmetric Deep Semantic Quantization for Image Retrieval
• Auto-Encoding Twin-Bottleneck Hashing
• Compact Hash Code Learning with Binary Deep Neural Network
• MILDNet: A Lightweight Single Scaled Deep Ranking Architecture
• Embarrassingly Simple Binary Representation Learning
• Mean Local Group Average Precision (mLGAP): A New Performance Metric for Hashing-based Retrieval
• SCH-GAN: Semi-supervised Cross-modal Hashing by Generative Adversarial Network
• Deep Priority Hashing
• Query-adaptive Image Retrieval by Deep Weighted Hashing
• Ranking-based Deep Cross-modal Hashing
• E${}^2$BoWs: An End-to-End Bag-of-Words Model via Deep Convolutional Neural Network
• Deep Triplet Quantization
• SADIH: Semantic-Aware DIscrete Hashing
• Hadamard Codebook Based Deep Hashing
• Two-pronged Strategy: Lightweight Augmented Graph Network Hashing for Scalable Image Retrieval
• FMHash: Deep Hashing of In-Air-Handwriting for User Identification
• Backdoor Attack on Hash-based Image Retrieval via Clean-label Data Poisoning
• Faster Person Re-Identification
• Deep Residual Hashing
• Learning A Deep ${\ell }_{\infty }$ Encoder for Hashing
• Cross-modal Deep Metric Learning with Multi-task Regularization
• Hierarchy Neighborhood Discriminative Hashing for An Unified View of Single-Label and Multi-Label Image retrieval
• Accurate Deep Representation Quantization with Gradient Snapping Layer for Similarity Search
• Dual Asymmetric Deep Hashing Learning
• Transductive Zero-Shot Hashing via Coarse-to-Fine Similarity Mining
• Prototype-supervised Adversarial Network for Targeted Attack of Deep Hashing
• Instance-weighted Central Similarity for Multi-label Image Retrieval
• Neural Audio Fingerprint for High-specific Audio Retrieval based on Contrastive Learning
• Targeted Attack for Deep Hashing based Retrieval
• Push for Quantization: Deep Fisher Hashing
• Weakly Supervised Deep Image Hashing through Tag Embeddings
• Simultaneous Compression and Quantization: A Joint Approach for Efficient Unsupervised Hashing
• Joint Cluster Unary Loss for Efficient Cross-Modal Hashing
• Deep Class-Wise Hashing: Semantics-Preserving Hashing via Class-wise Loss
• Learning Effective Binary Visual Representations with Deep Networks
• Triplet-Based Deep Hashing Network for Cross-Modal Retrieval
• Fast Supervised Discrete Hashing
• A survey on deep hashing for image retrieval
• Neural networks behave as hash encoders: An empirical study
• Discrete Few-Shot Learning for Pan Privacy
• Deep Supervised Hashing with Triplet Labels
• Rescuing Deep Hashing from Dead Bits Problem
• Deep Discrete Hashing with Self-supervised Pairwise Labels
• Deep Momentum Uncertainty Hashing
• Learning Robust Hash Codes for Multiple Instance Image Retrieval
• Scalable Discrete Supervised Hash Learning with Asymmetric Matrix Factorization
• Deep Supervised Hashing leveraging Quadratic Spherical Mutual Information for Content-based Image Retrieval
• Neurons Merging Layer: Towards Progressive Redundancy Reduction for Deep Supervised Hashing
• A Review of Computer Vision Methods in Network Security
• Dual-level Semantic Transfer Deep Hashing for Efficient Social Image Retrieval
• Deep Reinforcement Learning with Label Embedding Reward for Supervised Image Hashing
• Modal-aware Features for Multimodal Hashing
• A Scalable Optimization Mechanism for Pairwise based Discrete Hashing
• Pairwise Teacher-Student Network for Semi-Supervised Hashing
• Deep Policy Hashing Network with Listwise Supervision
• Supervised Discrete Hashing with Relaxation
• Query-Adaptive Hash Code Ranking for Large-Scale Multi-View Visual Search
• Fast Binary Embedding via Circulant Downsampled Matrix -- A Data-Independent Approach
• PDH : Probabilistic deep hashing based on MAP estimation of Hamming distance
• Binary Stochastic Representations for Large Multi-class Classification
• Online Supervised Hashing for Ever-Growing Datasets
• Learning a Complete Image Indexing Pipeline
• Hashing with Binary Matrix Pursuit
• Bilinear Supervised Hashing Based on 2D Image Features
• Seeker: Real-Time Interactive Search
• Multiple Code Hashing for Efficient Image Retrieval
• Deep Semantic Hashing with Generative Adversarial Networks
• Improving Deep Binary Embedding Networks by Order-aware Reweighting of Triplets
• Improved Deep Hashing with Soft Pairwise Similarity for Multi-label Image Retrieval
• Regularizing Deep Hashing Networks Using GAN Generated Fake Images
• Triplet-based Deep Similarity Learning for Person Re-Identification
• Self-supervised asymmetric deep hashing with margin-scalable constraint
• Unsupervised Triplet Hashing for Fast Image Retrieval
• Informative and Representative Triplet Selection for Multilabel Remote Sensing Image Retrieval
• Open-Set Representation Learning through Combinatorial Embedding
• Deep Learning to Ternary Hash Codes by Continuation
• ForestHash: Semantic Hashing With Shallow Random Forests and Tiny Convolutional Networks
• Improved Deep Classwise Hashing With Centers Similarity Learning for Image Retrieval
• Streaming Binary Sketching based on Subspace Tracking and Diagonal Uniformization