Research map / Architectures and efficiency / Neural architecture search
Efficient Joint-Dimensional Search with Solution Space Regularization for Real-Time Semantic Segmentation
Peng Ye, Baopu Li, Tao Chen, Jiayuan Fan, Zhen Mei, Lin Chen, Chongyan Zuo, Qinghua Chi, Wanli Ouyang
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 SSR loss (ablation) | progressive shrinking search · Cityscapes val dilation search | mIoU (%): 66.95 -> 68.06 | Improved |
| add SSR loss to PDARTS (ablation) | PDARTS · Cityscapes val dilation search | mIoU (%): 66.09 -> 66.36 | Improved |
| add SSR loss to SGAS (ablation) | SGAS · Cityscapes val dilation search | mIoU (%): 66.59 -> 67.25 | Improved |
| use level-based progressive shrinking instead of SGAS-style (ablation) | Ours without SSR · Cityscapes val dilation search | mIoU (%): 66.59 -> 66.95 | Improved |
| use SSR loss instead of L1 loss (ablation) | progressive shrinking · Cityscapes val dilation search | mIoU (%): 65.47 -> 68.06 | Improved |
| use SSR loss instead of L2 loss (ablation) | progressive shrinking · Cityscapes val dilation search | mIoU (%): 65.81 -> 68.06 | Improved |
| use our search scheme instead of DARTS (ablation) | dilation search · Cityscapes val | mIoU (%): 65.81 -> 68.06 | Improved |
| add channel, spatial and depth dimensions to dilation-only search under FLOPs constraint (ablation) | joint search · Cityscapes val, 7.8G FLOPs | mIoU (%): 64.63 -> 67.49 | Improved |
| add channel, spatial and depth dimensions to dilation-only search without FLOPs constraint (ablation) | joint search · Cityscapes val | mIoU (%): 68.04 -> 68.06 | No clear effect |
| add manual dilation arrangement to baseline1 (ablation) | baseline1 · Cityscapes val | mIoU (%): 59.38 -> 64.84 | Improved |
| add aggregation block in retraining (ablation) | searched model · Cityscapes val | mIoU (%): 68.70 -> 69.55 | Improved |
| add common training skills in retraining (ablation) | searched model · Cityscapes val | mIoU (%): 69.55 -> 73.67 | Improved |
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.
- DARTS — DARTS-: Robustly Stepping out of Performance Collapse Without Indicators (ICLR 2021) · better on 1 of 1 shared result
- FairDARTS — Fair DARTS: Eliminating Unfair Advantages in Differentiable Architecture Search (ECCV 2020) · 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
- FasterSeg: Searching for Faster Real-time Semantic Segmentation (ICLR 2020 · similarity 0.91)
- Graph-Guided Architecture Search for Real-Time Semantic Segmentation (CVPR 2020 · similarity 0.89)
- Fast Neural Architecture Search of Compact Semantic Segmentation Models via Auxiliary Cells (CVPR 2019 · similarity 0.88)
- Auto-DeepLab: Hierarchical Neural Architecture Search for Semantic Image Segmentation (CVPR 2019 · similarity 0.87)
- Customizable Architecture Search for Semantic Segmentation (CVPR 2019 · similarity 0.87)
- SasWOT: Real-Time Semantic Segmentation Architecture Search WithOut Training (AAAI 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.
