Research map / Architectures and efficiency / Neural architecture search
Real-Time Semantic Segmentation via Auto Depth, Downsampling Joint Decision and Feature Aggregation
Peng Sun, Jiaxiang Wu, Songyuan Li, Peiwen Lin, Junzhou Huang, Xi Li
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 searched aggregation cell (ablation) | AutoRTNet-A · Cityscapes val (ablation) | mIoU (%): 69.9 -> 72.9 | Improved |
| use searched aggregation cell instead of random (ablation) | AutoRTNet-A · Cityscapes val (ablation) | mIoU (%): 71.4 -> 72.9 | Improved |
| use gradient-based joint search instead of random search (ablation) | AutoRTNet · Cityscapes val (ablation) | mIoU (%): 66.7 -> 72.9 | Improved |
| use searched downsampling positions (1,3,7) instead of random (1,5,10) (ablation) | AutoRTNet-A · Cityscapes val (ablation) | mIoU (%): 71.6 -> 72.9 | Improved |
| use searched downsampling positions (1,3,7) instead of random (1,2,4) (ablation) | AutoRTNet-A · Cityscapes val (ablation) | mIoU (%): 68.4 -> 72.9 | Mixed |
| vary initial hyper-cell numbers and random seeds (ablation) | hyper-cell · Cityscapes val (ablation) | mIoU (%): 72.5 -> 73.0 | No clear effect |
| increase latency weight gamma from 0.001 to 0.1 (ablation) | AutoRTNet · CamVid test | mIoU (%): 73.3 -> 68.6 | Hurt |
| AutoRTNet-A vs CAS on Cityscapes test | AutoRTNet-A · Cityscapes test | mIoU (%): 70.5 -> 72.2 | Improved |
| AutoRTNet-A vs CAS on CamVid | AutoRTNet-A · CamVid test (720x960) | mIoU (%): 71.2 -> 73.5 | Improved |
Closest papers by meaning
- Auto-DeepLab: Hierarchical Neural Architecture Search for Semantic Image Segmentation (CVPR 2019 · similarity 0.89)
- DFANet: Deep Feature Aggregation for Real-Time Semantic Segmentation (CVPR 2019 · similarity 0.88)
- BiSeNet V2: Bilateral Network with Guided Aggregation for Real-Time Semantic Segmentation (IJCV 2021 · similarity 0.88)
- Graph-Guided Architecture Search for Real-Time Semantic Segmentation (CVPR 2020 · similarity 0.87)
- Customizable Architecture Search for Semantic Segmentation (CVPR 2019 · similarity 0.87)
- SCTNet: Single-Branch CNN with Transformer Semantic Information for Real-Time Segmentation (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.
