Research map / Detection and segmentation / CNN features
Bi-Directional Cascade Network for Perceptual Edge Detection
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use BDCN with multi-scale inputs instead of CED with multi-scale inputs | BSDS500 test · BSDS500 test | ODS: .815 -> .828 | Helped |
| Use BDCN with multi-scale inputs instead of RCF with multi-scale inputs | BSDS500 test · BSDS500 test | ODS: .811 -> .828 | Helped |
| Use BDCN on RGB instead of AMH-Net-ResNet50 on RGB | NYUDv2 RGB · NYUDv2 RGB | ODS: .744 -> .748 | Helped |
| Use BDCN instead of RCF with multi-scale inputs | Multicue boundary · Multicue boundary | ODS: .825 -> .838 | Helped |
| Use BDCN instead of RCF with multi-scale inputs | Multicue edge · Multicue edge | ODS: .860 -> .894 | Helped |
| Add SEM to the baseline (ablation) | BSDS500 validation BDCN · BSDS500 validation | ODS: .7681 -> .7748 | Helped |
| Add bidirectional S2D+D2S cascade to the baseline (ablation) | BSDS500 validation BDCN without SEM · BSDS500 validation | ODS: .7681 -> .7762 | Helped |
| Combine S2D and D2S cascades rather than using either direction alone (ablation) | BSDS500 validation · BSDS500 validation | ODS: .7683 -> .7762; .7710 -> .7762 | Helped |
| Increase SEM convolution count from K=0 to K=3 (ablation) | BSDS500 validation · BSDS500 validation | ODS: .7728 -> .7748 | Helped |
| Increase SEM convolution count from K=3 to K=4 (ablation) | BSDS500 validation · BSDS500 validation | ODS: .7748 -> .7745 | Hurt |
| Increase SEM dilation factor from r0=0 to r0=4 (ablation) | BSDS500 validation · BSDS500 validation | ODS: .7720 -> .7748 | Helped |
| Increase SEM dilation factor from r0=4 to r0=8 (ablation) | BSDS500 validation · BSDS500 validation | ODS: .7748 -> .7742 | Hurt |
Closest papers by meaning
- Feature Pyramid Networks for Object Detection (CVPR 2017 · similarity 0.87)
- ScaleNet - Improve CNNs through Recursively Rescaling Objects (AAAI 2020 · similarity 0.87)
- Multi-Scale Bidirectional FCN for Object Skeleton Extraction (AAAI 2018 · similarity 0.86)
- Data-Driven Neuron Allocation for Scale Aggregation Networks (CVPR 2019 · similarity 0.86)
- An Analysis of Scale Invariance in Object Detection SNIP (CVPR 2018 · similarity 0.86)
- 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.