Research map / Detection and segmentation / CNN features
Bayesian Adaptive Superpixel Segmentation
Roy Uziel, Meitar Ronen, Oren Freifeld
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 Potts term (spatial coherence prior) (ablation) | BASS · BSD (SP#=600) | EV: 0.906 -> 0.923 | Improved |
| add Potts term (ablation) | BASS · BSD (SP#=950) | Rec: 0.858 -> 0.849 | Hurt |
| add splits and merges (adaptive K) (ablation) | BASS · BSD (SP#=350) | EV: 0.884 -> 0.910 | Improved |
| add splits and merges (ablation) | BASS · BSD (SP#=950) | Rec: 0.843 -> 0.849 | Improved |
| add both Potts term and splits/merges (ablation) | BASS · BSD (SP#=600) | Rec: 0.887 -> 0.817 | Hurt |
| add both Potts term and splits/merges (ablation) | BASS · BSD (SP#=600) | EV: 0.909 -> 0.923 | Improved |
| add topology/connectivity, Potts and splits/merges to plain DPGMM (ablation) | DPGMM · BSD (SP#=950) | EV: 0.764 -> 0.931 | Improved |
| add topology/connectivity, Potts and splits/merges to plain DPGMM (ablation) | DPGMM · BSD (SP#=600) | Rec: 0.574 -> 0.817 | Improved |
| use BASS instead of reSEEDS superpixels for face detection | reSEEDS · Face detection (K=1100) | success rate: 25.733% -> 45.469% | Improved |
| use BASS instead of FSCSP superpixels for face detection | FSCSP · Face detection (K=550) | success rate: 14.209% -> 32.123% | Improved |
| add Bilateral Inception with SLIC superpixels to DeepLab | DeepLab · Pascal VOC 2012 test | mean IoU: 68.9 -> 74.42 | Improved |
| replace SLIC with BASS superpixels in DeepLab+BI | DeepLab+BI · Pascal VOC 2012 test | mean IoU: 74.42 -> 74.68 | No clear effect |
Closest papers by meaning
- Bayesian-Inspired Space-Time Superpixels (ICCV 2025 · similarity 0.87)
- Superpixel Segmentation With Fully Convolutional Networks (CVPR 2020 · similarity 0.87)
- Extract and Merge: Superpixel Segmentation with Regional Attributes (ECCV 2020 · similarity 0.86)
- Differentiable Laplacian Matrix Guided Superpixel Segmentation (CVPR 2026 · similarity 0.86)
- Hierarchical Histogram Threshold Segmentation - Auto-terminating High-detail Oversegmentation (CVPR 2024 · similarity 0.84)
- Learning Invariant Inter-pixel Correlations for Superpixel Generation (AAAI 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.
