Seeking commonness and inconsistencies: A jointly smoothed approach to multi-view subspace clustering

Multi-view subspace clustering aims to discover the hidden structures from multiple views for robust clustering, and has been attracting considerable attention in recent years. Despite significant progress, most of previous multi-view algorithms are still faced with two limitations. First, they usually focus on consistency (or commonness) views, yet often lack ability capture cross-view inconsistencies representations. Second, many them overlook local cannot jointly leverage enhance representation learning. To address these limitations, this paper, we propose a smoothed (JSMC) approach. Specifically, simultaneously incorporate commonness into The view-consensus grouping effect is presented exploit regularize view-commonness representation, which further associated low-rank constraint via nuclear norm strengthen its cluster structure. Thus inconsistencies, effect, seamlessly incorporated unified objective function, upon an alternating optimization algorithm performed achieve clustering. Experimental results variety real-world datasets confirm superiority our Code available: https://github.com/huangdonghere/JSMC.