Research map / Label-efficient and robust learning / Self-supervised and contrastive
Semantically-Guided Representation Learning for Self-Supervised Monocular Depth
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 fixed semantic guidance to PackNet | PackNet · KITTI depth up to 80m | Abs Rel: 0.108 -> 0.102 | Helped |
| Add fixed semantic guidance to PackNet | PackNet · NuScenes validation | Abs Rel: 0.187 -> 0.181 | Helped |
| Add semantic guidance (ablation) | ResNet-18 depth network · KITTI depth up to 80m | Abs Rel: 0.120 -> 0.117 | Helped |
| Add semantic guidance (ablation) | ResNet-50 depth network · KITTI depth up to 80m | Abs Rel: 0.117 -> 0.113 | Helped |
| Add semantic guidance (ablation) | PackNet · KITTI depth up to 80m | Abs Rel: 0.108 -> 0.103 | Helped |
| Add two-stage training after semantic guidance (ablation) | PackNet with semantic guidance · KITTI depth up to 80m | Class-average Abs Rel: 0.121 -> 0.117 | Helped |
| Add two-stage training after semantic guidance (ablation) | PackNet with semantic guidance · KITTI depth up to 80m | Abs Rel: 0.103 -> 0.102 | No clear effect |
| Pretrain the semantic guidance network on CityScapes rather than ImageNet (ablation) | Proposed method, depth network fine-tuned only · KITTI depth up to 80m | Abs Rel: 0.197 -> 0.102 | Helped |
| Fine-tune the semantic guidance network together with the depth network (ablation) | Proposed method with ImageNet-pretrained semantic network · KITTI depth up to 80m | Abs Rel: 0.197 -> 0.116 | Helped |
| Fine-tune the semantic guidance network together with the depth network (ablation) | Proposed method with half-CityScapes-pretrained semantic network · KITTI depth up to 80m | Abs Rel: 0.104 -> 0.109 | Hurt |
| Fine-tune the semantic guidance network together with the depth network (ablation) | Proposed method with CityScapes-pretrained semantic network · KITTI depth up to 80m | Abs Rel: 0.102 -> 0.107 | Hurt |
Closest papers by meaning
- Three Ways To Improve Semantic Segmentation With Self-Supervised Depth Estimation (CVPR 2021 · similarity 0.91)
- Improved Monocular Depth Prediction Using Distance Transform Over Pre-semantic Contours with Self-supervised Neural Networks (CVPR 2025 · similarity 0.89)
- SQLdepth: Generalizable Self-Supervised Fine-Structured Monocular Depth Estimation (AAAI 2024 · similarity 0.87)
- Unsupervised Semantic Segmentation Through Depth-Guided Feature Correlation and Sampling (CVPR 2024 · similarity 0.87)
- Mining Supervision for Dynamic Regions in Self-Supervised Monocular Depth Estimation (CVPR 2024 · similarity 0.87)
- From-Ground-To-Objects: Coarse-to-Fine Self-supervised Monocular Depth Estimation of Dynamic Objects with Ground Contact Prior (CVPR 2024 · similarity 0.87)
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.