Research map / Detection and segmentation / CNN features
Fully Convolutional Neural Networks with Full-Scale-Features for Semantic Segmentation
Read the paperCheck 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 |
|---|---|---|---|
| add FFN with multi-layer concatenation fusion (d) | DeepLab-LargeFOV · PASCAL VOC 2012 val | mean IoU (%): 67.7% -> 69.7% | Helped |
| add FFN with multi-layer concatenation fusion (d) | DeepLab-LargeFOV · PASCAL VOC 2012 test | mean IoU (%): 70.3% -> 71.3% | Helped |
| add FFN with concatenation fusion (d) | FCN-8s · PASCAL VOC 2012 val (736 images) | mean IoU (%): 62.7% -> 64.5% | Helped |
| fuse FFN and PFG features by element-wise sum instead of concatenation (ablation) | FCN-8s-FFN · PASCAL VOC 2012 val (736 images) | mean IoU (%): 64.2% -> 57.4% | Hurt |
| add re-fusion of FFN(m) with m (multi-layer fusion d over b) (ablation) | FCN-8s-FFN · PASCAL VOC 2012 val (736 images) | mean IoU (%): 64.2% -> 64.5% | Helped |
| add re-fusion of FFN(m) with m (d over b) (ablation) | DeepLab-LargeFOV-FFN · PASCAL VOC 2012 val | mean IoU (%): 69.3% -> 69.7% | Helped |
| add auxiliary 7x7 grid supervision on FFN output (ablation) | DeepLab-LargeFOV-FFN · PASCAL VOC 2012 val | mean IoU (%): 67.7% -> 67.4% | Hurt |
| replace ParseNet global pooling with FFN grid-based full-scale features | DeepLab-LargeFOV · PASCAL VOC 2012 test | mean IoU (%): 69.8% -> 71.3% | Helped |
| FFN full-scale features versus multi-scale (MSc) features | DeepLab-LargeFOV · PASCAL VOC 2012 test | mean IoU (%): 71.6% vs 71.3% | No clear effect |
Closest papers by meaning
- Learning Fully Dense Neural Networks for Image Semantic Segmentation (AAAI 2019 · similarity 0.90)
- Semantic Segmentation With Boundary Neural Fields (CVPR 2016 · similarity 0.89)
- Context Encoding for Semantic Segmentation (CVPR 2018 · similarity 0.89)
- Attention to Scale: Scale-Aware Semantic Image Segmentation (CVPR 2016 · similarity 0.88)
- Fast Neural Architecture Search of Compact Semantic Segmentation Models via Auxiliary Cells (CVPR 2019 · similarity 0.87)
- Unsupervised Semantic Segmentation Through Depth-Guided Feature Correlation and Sampling (CVPR 2024 · similarity 0.86)
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.