Research map / Detection and segmentation / Open-set and few-shot detection
ROAD: Reality Oriented Adaptation for Semantic Segmentation of Urban Scenes
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 target guided distillation (ablation) | DeepLab V2 (VGG-16) · Cityscapes val (GTAV) | mean IoU: 21.9 -> 31.1 | Helped |
| Add spatial-aware adaptation (ablation) | DeepLab V2 (VGG-16) · Cityscapes val (GTAV) | mean IoU: 21.9 -> 32.6 | Helped |
| Add both target guided distillation and spatial-aware adaptation (ablation) | DeepLab V2 (VGG-16) · Cityscapes val (GTAV) | mean IoU: 21.9 -> 35.9 | Helped |
| Add target guided distillation (ablation) | PSPNet (ResNet-101) · Cityscapes val (GTAV) | mean IoU: 27.8 -> 36.8 | Helped |
| Add spatial-aware adaptation (ablation) | PSPNet (ResNet-101) · Cityscapes val (GTAV) | mean IoU: 27.8 -> 37.8 | Helped |
| Add both target guided distillation and spatial-aware adaptation (ablation) | PSPNet (ResNet-101) · Cityscapes val (GTAV) | mean IoU: 27.8 -> 39.4 | Helped |
| Use ROAD-Net instead of Curriculum | VGG-16 segmentation model · Cityscapes val (GTAV) | mean IoU: 28.9 -> 35.9 | Helped |
| Use ROAD-Net instead of Curriculum | VGG-16 segmentation model · Cityscapes val (SYNTHIA) | mean IoU: 29.0 -> 36.2 | Helped |
| Replace target real images with synthetic source images for distillation (ablation) | DeepLab V2 target guided distillation · Cityscapes val (GTAV) | mean IoU | Hurt |
| Freeze the first few convolutional layers instead of using target guided distillation (ablation) | DeepLab V2 · Cityscapes val (GTAV) | mean IoU | Hurt |
| Use spatial partitions rather than global alignment (ablation) | DeepLab V2 spatial-aware adaptation · Cityscapes val (GTAV) | mean IoU | Helped |
| Increase partitioning from 2 × 1 to 2 × 2 (ablation) | DeepLab V2 spatial-aware adaptation · Cityscapes val (GTAV) | mean IoU | No clear effect |
Closest papers by meaning
- Learning Semantic Segmentation From Synthetic Data: A Geometrically Guided Input-Output Adaptation Approach (CVPR 2019 · similarity 0.90)
- The SYNTHIA Dataset: A Large Collection of Synthetic Images for Semantic Segmentation of Urban Scenes (CVPR 2016 · similarity 0.90)
- Learning From Synthetic Data: Addressing Domain Shift for Semantic Segmentation (CVPR 2018 · similarity 0.89)
- Not All Areas Are Equal: Transfer Learning for Semantic Segmentation via Hierarchical Region Selection (CVPR 2019 · similarity 0.89)
- Transferring and Regularizing Prediction for Semantic Segmentation (CVPR 2020 · similarity 0.89)
- MRFP: Learning Generalizable Semantic Segmentation from Sim-2-Real with Multi-Resolution Feature Perturbation (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.