Research map / Vision-language and multimodal / Human and embodied modeling
Spatial Structure and Selective Text Jointly Facilitate Image Clustering
Zizheng Jiu, Feijiang Li, Jieting Wang, Yuhua Qian, Lu Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add textual modality to datasets with low compactness (tau<0.33) (ablation) | SATC · Datasets 1 and 2 | ACC: 89.8 -> 60.7; 85.8 -> 59.9 | Hurt |
| add textual modality to datasets with high compactness (tau>0.33) (ablation) | SATC · Datasets 13, 15, 18 | ACC: 60.6 -> 67.1; 52.2 -> 65.2; 47.4 -> 52.0 | Improved |
| use compactness-based textual selector with threshold 0.33 (ablation) | SATC · 18 datasets | ACC | Improved |
| add GAT spatial branch on ResNet-50 patch features (ablation) | SATC · 18 datasets avg | ACC: 69.1 -> 73.7 | Improved |
| add ResNet-50 node features over raw CLIP (ablation) | SATC · 18 datasets avg | ACC: 60.8 -> 69.1 | Improved |
| use GAT instead of GCN (ablation) | SATC · 18 datasets avg | ACC: 71.9 -> 73.7 | Improved |
| use GAT instead of Transformer (ablation) | SATC · 18 datasets avg | ACC: 70.4 -> 73.7 | Improved |
| combine visual and spatial features via mutual distillation (ablation) | SATC · 18 datasets avg | ACC/NMI/ARI | Improved |
| use SATC instead of TAC on ImageNet-Dogs | SATC · ImageNet-Dogs | ACC: 84.4 -> 91.4 | Improved |
| vary consistency and entropy loss weights lambda1, lambda2 (ablation) | SATC · CIFAR-10, ImageNet-10 | ACC | No clear effect |
| add spatial GAT on Dataset 13 (GCN vs GAT) (ablation) | SATC · Dataset 13 | ACC: 69.7 (GCN) -> 67.1 (GAT) | Hurt |
Closest papers by meaning
- Language-Assisted Image Clustering Guided by Discriminative Relational Signals and Adaptive Semantic Centers (NeurIPS 2026 · similarity 0.89)
- Image Clustering via the Principle of Rate Reduction in the Age of Pretrained Models (ICLR 2024 · similarity 0.88)
- Image Clustering Conditioned on Text Criteria (ICLR 2024 · similarity 0.87)
- Image Clustering with External Guidance (ICML 2024 · similarity 0.87)
- Multi-Modal Proxy Learning Towards Personalized Visual Multiple Clustering (CVPR 2024 · similarity 0.86)
- Text Grouping Adapter: Adapting Pre-trained Text Detector for Layout Analysis (CVPR 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
