Discriminant Cross-modal Hashing Algorithm Based On Semanic Coupling Correlation

With the proliferation of the Internet and user-generated content,as well as the growing prevalence of cameras,mobile phones,and social media,huge amounts of multimedia data are being produced,forming a unique kind of big data.Multi-modal is one of the important characteristics of big data.With the continuous development of the era of big data,data present in a multi-modal mixed form,and each modal data represent different aspects of the same instance.How to realize the interaction between different modes of retrieval has become a hot research topic.Cross-modal retrieval based on hashing has attracted much attention for its low storage cost and fast query speed.The core problem of cross-modal hash learning is how to efficiently share the semantic space of different modal data.Most of algorithms neglect the semantic discrimination of feature representation in the process of embedding the multi-modal data into the shared space,which leads to weaken the classification discrimination of the hash codes from different classes and reduce accuracy and robustness of the nearest neighbor search.In this paper,a linear discriminative cross-modal hashing learning algorithm with coupled semantic correlation is proposed.This algorithm integrates linear discriminative classifier and maximizing the correlation between cross-modals in the objective function of the model,and learns the compact hash code representation with strong classification discrimination in the shared Hamming space.To verify the effectiveness of our method,the algorithm is compared with the current relevant algorithms on four benchmark datasets(Wiki,LabelMe,Mirflickr,NUS-WIDE),the experimental results show that the proposed method has obvious advantages on the retrieval accuracy and the computational efficiency.