Research map / 3D vision / Depth and stereo
Towards Scene Understanding: Unsupervised Monocular Depth Estimation With Semantic-Aware Representation
Check 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 semantic segmentation training data (ablation) | SceneNet · KITTI Eigen split, 80m cap | Abs Rel: 0.128 -> 0.118 | Helped |
| Add semantic segmentation training data (ablation) | SceneNet · KITTI Eigen split, 50m cap | Abs Rel: 0.122 -> 0.112 | Helped |
| Use horizontal flipping to predict the right-view disparity from its aligned image (ablation) | SceneNet · KITTI split | Abs Rel: 0.116 -> 0.112 | Helped |
| Use a unified decoder rather than separate task decoders (ablation) | SceneNet · KITTI split | Abs Rel: 0.112 -> 0.111 | Helped |
| Condition the unified decoder on task identity (ablation) | SceneNet · KITTI split | Abs Rel: 0.111 -> 0.104 | Helped |
| Add left-right semantic consistency loss (ablation) | SceneNet · KITTI split | Abs Rel: 0.104 -> 0.104 | No clear effect |
| Add semantics-guided disparity smoothness loss (ablation) | SceneNet · KITTI split | Abs Rel: 0.104 -> 0.102 | Helped |
| Share geometric and semantic information through SceneNet (ablation) | SceneNet · KITTI split | mIOU: 33.83% -> 47.87% | Helped |
Closest papers by meaning
- SDC-Depth: Semantic Divide-and-Conquer Network for Monocular Depth Estimation (CVPR 2020 · similarity 0.92)
- Unsupervised Semantic Segmentation Through Depth-Guided Feature Correlation and Sampling (CVPR 2024 · similarity 0.90)
- Know Your Neighbors: Improving Single-View Reconstruction via Spatial Vision-Language Reasoning (CVPR 2024 · similarity 0.89)
- A Unified Framework for 3D Scene Understanding (NeurIPS 2024 · similarity 0.89)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 2024 · similarity 0.88)
- RefineAny3D: Depth Refinement as Semantic Alignment for Monocular 3D Detection (NeurIPS 2026 · 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.