Research map / Label-efficient and robust learning / Self-supervised and contrastive
Multi-Frame Self-Supervised Depth With Transformers
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 |
|---|---|---|---|
| replace SAD or SSIM cost-volume matching with cross-attention matching and high-response decoding (ablation) | DepthFormer · KITTI | AbsRel: 0.647 and 0.632 -> 0.264 | Helped |
| add context adjustment to high-response depth (ablation) | DepthFormer · KITTI | AbsRel: 0.264 -> 0.167 | Helped |
| jointly decode multi-frame cross-attention and single-frame features (ablation) | DepthFormer · KITTI | AbsRel: 0.167 -> 0.090 | Helped |
| remove cross-attention (ablation) | DepthFormer · KITTI | AbsRel: 0.090 -> 0.099 | Hurt |
| remove self-supervision losses on the cross-attention depth estimates (ablation) | DepthFormer · KITTI | AbsRel: 0.090 -> 0.103 | Hurt |
| remove self-attention from context features (ablation) | DepthFormer · KITTI | AbsRel: 0.090 -> 0.094 | Hurt |
| reduce depth bins to 16 (ablation) | DepthFormer · KITTI | AbsRel: 0.090 -> 0.101 | Hurt |
| use 48 depth bins instead of the full configuration (ablation) | DepthFormer · KITTI | AbsRel: 0.090 -> 0.095 | Hurt |
| use 96 depth bins instead of the full configuration (ablation) | DepthFormer · KITTI | AbsRel: 0.090 -> 0.092 | Hurt |
| reduce attention channels to 32 (ablation) | DepthFormer · KITTI | AbsRel: 0.090 -> 0.104 | Hurt |
| reduce attention channels to 48 (ablation) | DepthFormer · KITTI | AbsRel: 0.090 -> 0.098 | Hurt |
| reduce attention channels to 96 (ablation) | DepthFormer · KITTI | AbsRel: 0.090 -> 0.093 | Hurt |
Closest papers by meaning
- The Temporal Opportunist: Self-Supervised Multi-Frame Monocular Depth (CVPR 2021 · similarity 0.92)
- Self-Supervised Monocular Trained Depth Estimation Using Self-Attention and Discrete Disparity Volume (CVPR 2020 · similarity 0.92)
- Deep Digging into the Generalization of Self-Supervised Monocular Depth Estimation (AAAI 2023 · similarity 0.91)
- ROIFormer: Semantic-Aware Region of Interest Transformer for Efficient Self-Supervised Monocular Depth Estimation (AAAI 2023 · similarity 0.90)
- Self-supervised surround-view depth estimation with volumetric feature fusion (NeurIPS 2022 · similarity 0.90)
- Mining Supervision for Dynamic Regions in Self-Supervised Monocular Depth Estimation (CVPR 2024 · similarity 0.89)
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.