Research map / Detection and segmentation / CNN features
RINDNet: Edge Detection for Discontinuity in Reflectance, Illumination, Normal and Depth
Mengyang Pu, Yaping Huang, Qingji Guan, Haibin Ling
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 Weight Layer (ablation) | RINDNet · BSDS-RIND (average) | ODS: 0.468 -> 0.486 | Improved |
| add Weight Layer (ablation) | RINDNet · BSDS-RIND (reflectance) | ODS: 0.422 -> 0.478 | Improved |
| add Attention Module (ablation) | RINDNet · BSDS-RIND (average) | ODS: 0.464 -> 0.486 | Improved |
| add Attention Module (ablation) | RINDNet · BSDS-RIND (average) | AP: 0.376 -> 0.432 | Improved |
| add Weight Layer on illumination edges (ablation) | RINDNet · BSDS-RIND (illumination) | AP: 0.180 -> 0.168 AP | No clear effect |
| use res1-3 for RE&IE and res5 for NE&DE instead of res1-3 for both (ablation) | RINDNet · BSDS-RIND (average) | ODS: 0.467 -> 0.486 | Improved |
| use res1-3 for RE&IE and res5 for NE&DE instead of swapped layers (ablation) | RINDNet · BSDS-RIND (average) | ODS: 0.452 -> 0.486 | Improved |
| use f_sp1-3 for RE&IE and f_sp1-5 for NE&DE instead of f_sp1-3 for both (ablation) | RINDNet · BSDS-RIND (average) | ODS: 0.472 -> 0.486 | Improved |
| add spatial cues f_sp (ablation) | RINDNet · BSDS-RIND (average) | ODS: 0.478 -> 0.486 | Improved |
| add second stream to decoders (ablation) | RINDNet · BSDS-RIND (average) | ODS: 0.457 -> 0.486 | Improved |
| share weights of second stream between NE and DE decoders (ablation) | RINDNet · BSDS-RIND (average) | ODS: 0.474 -> 0.486 | Improved |
| extend generic detectors with four branches (DeepLabV3+, DOOBNet, OFNet) | baselines · BSDS-RIND (average) | ODS: 0.446 to 0.453 ODS | No clear effect |
Later papers that compared with this method
1 later accepted paper reports RINDNet as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- MixerCSeg: An Efficient Mixer Architecture for Crack Segmentation via Decoupled Mamba Attention (CVPR 2026 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- BDCN — BDCN: Bi-Directional Cascade Network for Perceptual Edge Detection (TPAMI 2020) · better on 1 of 1 shared result
- CASENet — CASENet: Deep Category-Aware Semantic Edge Detection (CVPR 2017) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Differential Geometry in Edge Detection: Accurate Estimation of Position, Orientation and Curvature (TPAMI 2018 · similarity 0.86)
- Unsupervised Learning of Edges (CVPR 2016 · similarity 0.85)
- BDCN: Bi-Directional Cascade Network for Perceptual Edge Detection (TPAMI 2020 · similarity 0.85)
- Accurate RGB-D Salient Object Detection via Collaborative Learning (ECCV 2020 · similarity 0.85)
- Harnessing Edge Information for Improved Robustness in Vision Transformers (AAAI 2024 · similarity 0.84)
- DiffusionEdge: Diffusion Probabilistic Model for Crisp Edge Detection (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.
