Research map / Detection and segmentation / Open-set and few-shot detection
PiCIE: Unsupervised Semantic Segmentation Using Invariance and Equivariance in Clustering
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use PiCIE with auxiliary over-clustering instead of Modified DC | COCO-All segmentation · COCO-All | Accuracy: 32.21 -> 49.99 | Helped |
| Use PiCIE with auxiliary over-clustering instead of Modified DC | Cityscapes segmentation · Cityscapes | mIoU: 7.06 -> 12.31 | Helped |
| Replace a parametric classifier with centroid prototypes (ablation) | PiCIE clustering · COCO-All | Accuracy: 34.35 -> 39.25 | Helped |
| Add photometric cross-view learning (ablation) | PiCIE with centroid prototypes · COCO-All | Accuracy: 39.25 -> 42.55 | Helped |
| Add geometric equivariance (ablation) | PiCIE with centroid prototypes · COCO-All | Accuracy: 39.25 -> 46.97 | Helped |
| Add both photometric invariance and geometric equivariance (ablation) | PiCIE with centroid prototypes · COCO-All | Accuracy: 46.97 -> 48.09 | Helped |
| Add auxiliary over-clustering (ablation) | PiCIE · COCO-All | Accuracy: 48.09 -> 49.99 | Helped |
| Remove cluster balancing (ablation) | PiCIE · COCO-All | Accuracy: 48.09 -> 40.56 | Hurt |
| Use a single clustering rather than two separately clustered views (ablation) | PiCIE · COCO-All | Accuracy: 48.09 -> 44.31 | Hurt |
| Replace cross-view clustering loss with MSE feature matching (ablation) | PiCIE · COCO-All | Accuracy: 48.09 -> 44.15 | Hurt |
| Remove photometric invariance (ablation) | PiCIE · COCO-All | Accuracy: 48.09 -> 41.70 | Hurt |
Closest papers by meaning
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · similarity 0.88)
- Integrating Low-Level Visual Cues for Enhanced Unsupervised Semantic Segmentation (AAAI 2025 · similarity 0.88)
- Segment Anything without Supervision (NeurIPS 2024 · similarity 0.87)
- Label-Efficient Few-Shot Semantic Segmentation with Unsupervised Meta-Training (AAAI 2024 · similarity 0.86)
- Learning Hierarchical Image Segmentation For Recognition and By Recognition (ICLR 2024 · similarity 0.86)
- Unsupervised Semantic Segmentation Through Depth-Guided Feature Correlation and Sampling (CVPR 2024 · similarity 0.86)
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.