Lune

CVPR Findings2026

Prototype and Sample Level Semantic Alignment for Incomplete Multi-View Clustering

Zhengzhong Zhu, Pei Zhou, Lanxi Bai, Jia Nie, Li Cheng, Shiquan Min, Jiangping Zhu

Abstract

Incomplete Multi-View Clustering (IMVC) is a significant challenge due to the frequent occurrence of missing values in multi-view datasets. Existing prototype-based approaches face two primary obstacles. First, the divergence in view-specific feature spaces leads to prototype-unaligned problem (PUP), where prototypes from different views fail to represent the same semantic concepts. Second, contrastive learning methods are often hindered by false negatives problem (FNP), where semantically related instances are mistakenly pushed apart, degrading the quality of the learned representations. To address these issues, we introduce Prototype and Sample Level Semantic Alignment for Incomplete Multi-View Clustering (PSAMC), a unified framework designed to resolve both challenges. PSAMC consists of two core modules. The first module, Prototype Semantic Alignment, uses optimal transport to ensure consistent alignment of prototypes across different views, effectively incorporating semantic similarities. This module addresses the PUP issue by ensuring that prototypes from different views accurately represent the same underlying semantic concepts. The second module, Sample Semantic Alignment, employs Dynamic Contrastive Loss to continuously filter false negatives. By adapting based on the model's state, this module ensures that only semantically meaningful samples are used for contrastive loss computation, thereby mitigating the impact of FNP. Extensive experiments demonstrate that PSAMC outperforms existing methods, offering a robust and effective solution to the challenges of incomplete multi-view clustering.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

  • DealMVC: Dual Contrastive Calibration for Multi-view Clustering cites Align then Fusion: Generalized Large-scale Multi-view Clustering with Anchor Matching Correspondences
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on18

  1. Bootstrap Your Own Latent - A New Approach to Self-Supervised LearningJean-Bastien Grill, Florian Strub, Florent Altché, Corentin Tallec et al.NeurIPS 2020 · 9,171 citations
  2. Graph Contrastive ClusteringHuasong Zhong, Jianlong Wu, Chong Chen, Jianqiang Huang et al.ICCV 2021 · 163 citations
  3. Deep Incomplete Multi-View Clustering via Mining Cluster ComplementarityJie Xu, Chao Li, Yazhou Ren, Liang Peng et al.AAAI 2022 · 149 citations
  4. Align then Fusion: Generalized Large-scale Multi-view Clustering with Anchor Matching CorrespondencesSiwei Wang, Xinwang Liu, Suyuan Liu, Jiaqi Jin et al.NeurIPS 2022 · 144 citations
  5. DealMVC: Dual Contrastive Calibration for Multi-view ClusteringXihong Yang, Jiaqi Jin, Siwei Wang, Ke Liang et al.ACM MM 2023 · 138 citations

Related papers

  1. Dual-Level Distribution Alignment for Deep Incomplete Multi-View ClusteringFujian Ren, Wenlan Chen, Lu Gao, Fei Guo et al.ACM MM 2025
  2. Robust Prototype Completion for Incomplete Multi-view ClusteringHonglin Yuan, Shiyun Lai, Xingfeng Li, Jian Dai et al.ACM MM 2024 · 29 citations
  3. Learning from Disjoint Views: A Contrastive Prototype Matching Network for Fully Incomplete Multi-View ClusteringYiming Wang, Qun Li, Dongxia Chang, Jie Wen et al.NeurIPS 2025
  4. Deep Incomplete Multi-View Clustering with Cross-View Partial Sample and Prototype AlignmentJiaqi Jin, Siwei Wang, Zhibin Dong, Xinwang Liu et al.CVPR 2023
  5. Robust Multi-view Clustering via Pseudo Label Guided Universum LearningZhenxi Wang, Zongyao Yin, Yujie Hou, Xianchuan YuACM MM 2025 · 1 citation
Dusk over the sea between two cliffs drawn in fine vertical lines