Research map / Detection and segmentation / CNN features
Piecewise Flat Embedding for Image Segmentation
Chaowei Fang, Zicheng Liao, Yizhou Yu
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 residual-based channel weighting (ablation) | soPFE0.8+DB · BSDS500 hierarchical segmentation | F_op (ODS): 0.438 -> 0.463 | Improved |
| use WSC initialization instead of GMM (with weighting) (ablation) | gwPFE0.8+DB · BSDS500 hierarchical segmentation | F_op (ODS): 0.440 -> 0.463 | Improved |
| vary initialization without channel weighting (GMM, hierarchical, WSC) (ablation) | oPFE0.8+DB · BSDS500 hierarchical segmentation | F_op (ODS): 0.435, 0.437, 0.438 | No clear effect |
| use WSC instead of random initialization with weighting (ablation) | wPFE0.8+DB · BSDS500 hierarchical segmentation | F_op (ODS): 0.445 -> 0.463 | Improved |
| use p=0.8 instead of p=0.5 (ablation) | swPFE+DB · BSDS500 hierarchical segmentation | F_op (ODS): 0.450 -> 0.463 | Improved |
| use p=0.8 instead of p=1 (ablation) | swPFE+DB · BSDS500 hierarchical segmentation | F_op (ODS): 0.462 -> 0.463 | No clear effect |
| replace Laplacian eigenmaps with piecewise flat embedding | clustering segmentation · BSDS500 clustering segmentation (NCut affinity, fixed K) | F_op: 54.4% higher | Improved |
| replace eigenmaps with PFE in gPb-owt-ucm | gPb-owt-ucm · BSDS500 hierarchical segmentation | F_op (ODS): 0.348 -> 0.364 | Improved |
| replace eigenmaps with PFE in gPb-owt-ucm | gPb-owt-ucm · BSDS500 hierarchical segmentation | VI (ODS): 1.690 -> 1.565 | Improved |
| replace eigenmaps with PFE in gPb-owt-ucm | gPb-owt-ucm · BSDS500 hierarchical segmentation | F_b (ODS): 0.729 -> 0.723 | Hurt |
| replace Laplacian eigenmaps with PFE | DB+eigenmaps · BSDS500 hierarchical segmentation | F_op (ODS): 0.445 -> 0.463 | Improved |
| replace Laplacian eigenmaps with PFE | COB+eigenmaps · PASCAL Context hierarchical segmentation | F_op (ODS): 0.350 -> 0.503 | Improved |
Closest papers by meaning
- Rethinking Decoders for Transformer-based Semantic Segmentation: A Compression Perspective (NeurIPS 2024 · similarity 0.85)
- Sparse Dimensionality Reduction Revisited (ICML 2024 · similarity 0.83)
- EFDTR: Learnable Elliptical Fourier Descriptor Transformer for Instance Segmentation (ICML 2025 · similarity 0.83)
- $\epsilon$-Seg: Sparsely Supervised Semantic Segmentation of Microscopy Data (NeurIPS 2025 · similarity 0.83)
- Fourier Sliced-Wasserstein Embedding for Multisets and Measures (ICLR 2025 · similarity 0.83)
- EAGLE: Eigen Aggregation Learning for Object-Centric Unsupervised Semantic Segmentation (CVPR 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.
