Adaptive multi-modal fusion hashing via Hadamard matrix

Hashing plays an important role in information retrieval, due to its low storage and high speed of processing. As effective multi-modal representation learning method, hashing has received particular attention. Most the existing methods adopt fixed weighting factors fuse multiple modalities for any query data, which cannot capture variation among different queries. Besides, there are too much hyper-parameters their models while it is time-consuming labor-intensive determine proper parameters. The limitations may significantly hinder promotion practical applications. In this paper, we propose a simple, yet method that inspired by Hadamard matrix. On one hand, our proposed involves very few flexible. other complementary between data semantic discrimination preserved well hash codes. Extensive experimental results on four benchmark datasets show framework achieves superior performance compared state-of-the-art methods.