WATCH: Two-Stage Discrete Cross-Media Hashing

Due to the explosive growth of multimedia data in recent years, cross-media hashing (CMH) approaches have recently received increasing attention. To learn the hash codes, most existing supervised CMH algorithms employ the strict binary label information, which has tiny gaps between the incorrect labels (0) and the true labels (1), increasing the risk of classification error. Besides, most existing CMH approaches are one-stage algorithms, in which the hash functions and binary codes can be learned synchronously, complicating the optimization. To avoid NP-hard optimization, which is the consequence of the discrete constraints of the objective function, many approaches utilize a relaxation strategy. However, this optimisation trick may cause large quantization errors. To address these issues, we present a novel t
W
o-st
A
ge discre
T
e
C
ross-media
H
ashing method based on a smooth matrix factorization and label relaxation, named WATCH. The proposed WATCH controls the margins adaptively by the novel label relaxation strategy. This innovation reduces the quantization error significantly. Besides, WATCH is a two-stage model. In stage 1, we employ a discrete smooth matrix factorization model. Then, the hash codes can be generated discretely, reducing the large quantization loss greatly. In stage 2, we adopt an effective hash function learning strategy, which produces a more effective hash function. Comprehensive experiments on three popular datasets demonstrate that WATCH achieves significantly better performance than several state-of-the-art approaches.