Research map / Detection and segmentation / CNN features
SegGBC: Justifiable Coarse-to-Fine Granular-Ball Computing for Enhancing Clustering Image Segmentation
Qianpeng Chong, Wenyi Zeng, Xiuxuan Shen, Jiajie Li, Qian Yin, Xin Zheng
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 |
|---|---|---|---|
| replace fuzzy set with intuitionistic fuzzy set (hesitation degree) (ablation) | SegGBC · NI 7 | SA (%): 80.86 -> 96.83 | Improved |
| replace fuzzy set with intuitionistic fuzzy set (ablation) | SegGBC · RSI 3 | SA (%): 74.65 -> 82.91 | Improved |
| add SCMGB semantic compactness criterion with stable region (ablation) | Intuitionistic fuzzy granular balls · NI 7 | SA (%): 72.80 -> 96.83 | Improved |
| add SCMGB and stable region to IFS granular balls (ablation) | SegGBC · RSI 3 | SA (%): 67.29 -> 82.91 | Improved |
| use stable-region radius with SCMGB instead of max radius (traditional GB) (ablation) | Granular balls · NI 7 | time cost (s): 3.16 -> 2.09 | Improved |
| use average radius instead of max radius (ablation) | Traditional granular balls · NI 7 | SA (%): 63.34 -> 69.20 | Improved |
| add SegGBs as plug-and-play front-end to DFKM | DFKM · NI 2 | SA: +38.22 points | Improved |
| add SegGBs front-end to PiCIE+H | PiCIE+H · COCO-Stuff | mIoU (%): 14.36 -> 21.04 | Improved |
| add SegGBs front-end to I2Clu | I2Clu · COCO-Stuff-3 | PixelAcc (%): 72.30 -> 73.69 | Improved |
| apply GB-based methods to remote sensing images | Ball k-means and MGNR · RSI 1-3 | 3-mIoU (%): 52.19 (Ball-kmeans) vs 57.79 (MK-SPCM) | Hurt |
Closest papers by meaning
- Structure-aware Granular-Ball based Information Bottleneck for Multi-modal Clustering (ICML 2026 · similarity 0.87)
- MC-H: Multi-Granularity Clustering with Hyperspherical Determinantal Point Process (NeurIPS 2026 · similarity 0.85)
- Segment and Recognize Anything at Any Granularity (ECCV 2024 · similarity 0.84)
- Uncertainty-aware Fine-tuning of Segmentation Foundation Models (NeurIPS 2024 · similarity 0.84)
- Multi-view Granular-ball Contrastive Clustering (AAAI 2025 · similarity 0.83)
- SAM-PARSER: Fine-Tuning SAM Efficiently by Parameter Space Reconstruction (AAAI 2024 · similarity 0.82)
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.
