Multi-view Clustering based on Doubly Stochastic Graph

Nian Wang, Zhigao Cui, Aihua Li, Rong Wang, Feiping Nie

Abstract: Highlights•We introduce doubly stochastic graph learning to filter out the noise in a graph and propose a new model (MCDSG) for multi-view clustering.•We innovatively propose a simple yet highly effective approach to optimize the doubly stochastic condition.•We propose a pipeline to add noise to the key locations of face images and obtain a noisy face dataset termed noisedORL.•The experiments show our MCDSG is more robust to noised data and achieves SOTA clustering performance on benchmarks.

External IDs:dblp:journals/sigpro/0001CL0026