Dual global information guidance for deep contrastive multi-modal clustering

Guoliang Zou, Shizhe Hu, Tongji Chen, Yunpeng Wu, Yangdong Ye

Abstract: Highlights•A dual global information guidance framework is proposed for multi-modal clustering.•A global feature-guided module is designed to eliminate irrelevant information.•A global cluster assignment-guided module is presented to improve partition accuracy.•Extensive experiments demonstrate the effectiveness of the proposed algorithm.