Research map / Detection and segmentation / CNN features
Simultaneous Edge Alignment and Learning
Zhiding Yu, Weiyang Liu, Yang Zou, Chen Feng, Srikumar Ramalingam, B. V. K. Vijaya Kumar, Jan Kautz
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 CASENet-S training labels with SEAL jointly aligned labels | CASENet-S · SBD test original labels (Thin) | mean MF (%): 73.8 -> 73.8 | No clear effect |
| replace CASENet-S training labels with SEAL jointly aligned labels | CASENet-S · SBD test original labels (Raw) | mean MF (%): 66.5 -> 70.3 | Improved |
| replace CASENet-S with SEAL | CASENet-S · SBD re-annotated test (Thin) | mean MF (%): 65.9 -> 67.0 | Improved |
| replace CASENet-S with SEAL | CASENet-S · SBD re-annotated test (Raw) | mean MF (%): 60.3 -> 64.4 | Improved |
| refine labels with dense-CRF before training (CASENet-C) | CASENet-S · SBD re-annotated test (Thin) | mean MF (%): 65.9 -> 64.8 | Hurt |
| replace CASENet-S with SEAL | CASENet-S · Cityscapes val (Raw) | mean MF (%): 63.3 -> 65.5 | Improved |
| replace CASENet-S with SEAL | CASENet-S · Cityscapes val (Thin) | mean MF (%): 69.1 -> 69.1 | No clear effect |
| replace prior state of the art with SEAL under prior benchmark | previous SOTA · SBD test, prior protocol | mean MF (%): 71.4 -> 74.4 | Improved |
| add Markov smoothness prior (lambda 0 -> 0.02) to alignment (ablation) | SEAL sigma 1,4 · SBD val 0.02 tolerance (Thin) | mean MF (%): 72.7 -> 73.4 | Improved |
| add Markov smoothness prior (lambda 0 -> 0.02) to alignment (ablation) | SEAL sigma 1,4 · SBD val 0.02 tolerance (Raw) | mean MF (%): 72.9 -> 70.0 | Hurt |
| add biased Gaussian kernel (sigma_x 4 -> 1) (ablation) | SEAL lambda 0 · SBD val 0.02 tolerance (Thin) | mean MF (%): 71.6 -> 72.7 | Improved |
| decrease sigma_y from 4 to 2 (ablation) | SEAL lambda 0.02 · SBD val 0.02 tolerance (Raw) | mean MF (%): 70.0 -> 68.4 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CASENet — CASENet: Deep Category-Aware Semantic Edge Detection (CVPR 2017) · better on 6 of 6 shared results
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
- Unsupervised Learning of Edges (CVPR 2016 · similarity 0.87)
- SAUGE: Taming SAM for Uncertainty-Aligned Multi-Granularity Edge Detection (AAAI 2025 · similarity 0.86)
- CASENet: Deep Category-Aware Semantic Edge Detection (CVPR 2017 · similarity 0.86)
- The Treasure Beneath Multiple Annotations: An Uncertainty-Aware Edge Detector (CVPR 2023 · similarity 0.86)
- RankED: Addressing Imbalance and Uncertainty in Edge Detection Using Ranking-based Losses (CVPR 2024 · similarity 0.85)
- Harnessing Edge Information for Improved Robustness in Vision Transformers (AAAI 2024 · similarity 0.85)
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.
