Deep Clustering With Hybrid-Grained Contrastive and Discriminative Learning

作     者：Huang, Dong Deng, Xiaozhi Chen, Ding-Hua Wen, Zihao Sun, Weijun Wang, Chang-Dong Lai, Jian-Huang 

作者机构：South China Agr Univ Coll Math & Informat Guangzhou 510642 Peoples R China Guangdong Univ Technol Sch Automat Guangzhou 510006 Peoples R China Sun Yat Sen Univ Sch Comp Sci & Engn Guangzhou 510275 Peoples R China 

出 版 物：《IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY》 (IEEE Trans Circuits Syst Video Technol)

年 卷 期：2024年第34卷第10期

页      面：9472-9483页

核心收录：

学科分类：0808[工学-电气工程] 08[工学] 

基　　金：NSFC [61976097, 62276277, U22A2095] Guangdong Project [2021JC06X667] Guangdong Forestry Science Data Center [2021B1212100004] Guangdong Provincial Key Laboratory of Intellectual Property and Big Data [2018B030322016] 

主　　题：Feature extraction Representation learning Semantics Clustering methods Clustering algorithms Prototypes Network architecture Data clustering image clustering deep clustering contrastive clustering granularity 

摘      要：Deep contrastive clustering has recently gained significant attention due to its advantageous ability to leverage the contrastive learning paradigm for joint representation learning and clustering. However, previous deep contrastive clustering approaches mostly focus on instance discrimination or cluster discrimination, which often overlook the rich semantic information latent in the vast intermediate levels of granularity between instances and clusters. Moreover, they are typically prone to utilizing relationships only within the same level of granularity, e.g., instance-instance relationships and cluster-cluster relationships, but frequently neglect the interactions between different granularity-levels. To tackle these issues, this paper presents a novel end-to-end deep contrastive clustering approach termed Deep Clustering with Hybrid-Grained Contrastive and Discriminative Learning (DCHL). Particularly, the instance-level contrastive learning and cluster-level contrastive learning are first formulated, where the cluster-level contrastive learning is further split into fine-grained and coarse-grained branches. To capture global dependencies, the cluster-level contrastiveness is explored on the coarse-grained cluster branch. Meanwhile, to capture hybrid-grained relationships, the dual-level instance-group discrimination learning is enforced between the instance branch and the fine-grained cluster branch, where the self instance-group discrimination and the cross instance-group discrimination are simultaneously optimized for enhancing the deep clustering performance. Experiments on five challenging image datasets confirm the superiority of DCHL over the state-of-the-art. Code available: https://***/dengxiaozhi/DCHL.