Deep embedded multi-view clustering with collaborative training

Multi-view clustering has attracted increasing attentions recently by utilizing information from multiple views. However, existing multi-view methods are either with high computation and space complexities, or lack of representation capability. To address these issues, we propose deep embedded collaborative training (DEMVC) in this paper. Firstly, the representations views learned individually autoencoders. Then, both consensus complementary taken into account a novel scheme is proposed. Concretely, feature cluster assignments all collaboratively. A new consistency strategy for centers initialization further developed to improve performance training. Experimental results on several popular datasets show that DEMVC achieves significant improvements over state-of-the-art methods.