Abstract:Cross-modal hashing has drawn considerable interest in multimodal retrieval due to the explosive growth of big data on multimedia. However, the existing methods mainly focus on unified hash codes learning and investigate the local geometric structure in the original space, resulting in low-discriminative power hash code of out-of-sample instances. To address this important problem, this paper is dedicated to investigate the hashing functions learning by considering the modality correlation preserving in the expected low-dimensional common space. A cross-modal hashing method based on supervised collective matrix factorization is proposed by taking intra-modality and inter-modality similarity preserving into account. For more flexible hashing functions, label information is embedded into the hashing functions learning procedure. Specifically, we explore the intra-modality similarity preserving in the expected low-dimensional common space. In addition, a supervised shrinking scheme is used to enhance the local geometric consistency in each modality. The proposed method learns unified hash codes as well as hashing functions for different modalities; the overall objective function, consisting of collective matrix factorization and intra- and inter-modality similarity embedding, is solved using an alternative optimization in an iterative scheme. Extensive experiments on three benchmark data sets demonstrate that the proposed method is more flexible to new coming data and can achieve superior performance to the state-of-the-art supervised cross-modal hashing approaches in most of the cases.

Kernelised Supervised Context Hashing

Nonlinear Discrete Cross-Modal Hashing for Visual-Textual Data

Efficient Discrete Supervised Hashing for Large-scale Cross-modal Retrieval

Supervised Coarse-to-Fine Semantic Hashing for Cross-Media Retrieval.

Supervised hashing with kernels

Kernel-Based Supervised Discrete Hashing For Image Retrieval

Image Retrieval via Balanced and Maximum Variance Deep Hashing.

Exploring Auxiliary Context: Discrete Semantic Transfer Hashing for Scalable Image Retrieval.

Deep Supervised Hashing with Information Loss.

Supervised Discrete Hashing with Mutual Linear Regression.

Natural Supervised Hashing

Graph-based Supervised Discrete Image Hashing

Discriminative Supervised Hashing for Cross-Modal similarity Search

Robust Supervised Hashing.

Deep Supervised Hashing Using Symmetric Relative Entropy.

Adaptive Object Retrieval with Kernel Reconstructive Hashing

Multi-Kernel Hashing With Semantic Correlation Maximization For Cross-Modal Retrieval

Binary Multidimensional Scaling for Hashing

Strongly Constrained Discrete Hashing

Supervised Intra- and Inter-Modality Similarity Preserving Hashing for Cross-Modal Retrieval.

Discriminative Codebook Hashing for Supervised Video Retrieval