Research map / Label-efficient and robust learning / Self-supervised and contrastive
Deep Digging into the Generalization of Self-Supervised Monocular Depth Estimation
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 |
|---|---|---|---|
| Use the CNN-Transformer hybrid with multi-level feature fusion instead of CNN or Transformer baselines | Self-supervised monocular depth estimation · KITTI Eigen test | Abs Rel: 0.111 -> 0.104 | Helped |
| Use the CNN-Transformer hybrid instead of the strongest listed baseline | KITTI-trained depth estimation · MVS | Abs Rel: 0.260 -> 0.240 | Helped |
| Use the CNN-Transformer hybrid instead of the strongest listed baseline | KITTI-trained depth estimation · RGBD | Abs Rel: 0.383 -> 0.363 | Helped |
| Use the CNN-Transformer hybrid instead of the strongest listed baseline | KITTI-trained depth estimation · Scenes11 | Abs Rel: 1.647 -> 1.511 | Helped |
| Use the CNN-Transformer hybrid instead of the strongest listed baseline | KITTI-trained depth estimation · SUN3D | Abs Rel: 0.289 -> 0.245 | Helped |
| Use the CNN-Transformer hybrid instead of the strongest listed baseline | KITTI-trained depth estimation · ETH3D | Abs Rel: 0.701 -> 0.668 | Helped |
| Add ACM to the DPT baseline without FFD (ablation) | MonoFormer · KITTI Eigen test | Abs Rel: 0.113 -> 0.113 | No clear effect |
| Add FFD to the DPT baseline without ACM (ablation) | MonoFormer · KITTI Eigen test | Abs Rel: 0.113 -> 0.112 | Helped |
| Add ACM and FFD together (ablation) | MonoFormer · KITTI Eigen test | Abs Rel: 0.113 -> 0.104 | Helped |
| Use 4 encoder and decoder layers rather than 2 (ablation) | MonoFormer · KITTI Eigen test | Abs Rel: 0.148 -> 0.104 | Helped |
| Use 4 encoder and decoder layers rather than 3 (ablation) | MonoFormer · KITTI Eigen test | Abs Rel: 0.112 -> 0.104 | Helped |
| Use 4 encoder and decoder layers rather than 5 (ablation) | MonoFormer · KITTI Eigen test | Abs Rel: 0.111 -> 0.104 | Helped |
Closest papers by meaning
- Multi-Frame Self-Supervised Depth With Transformers (CVPR 2022 · similarity 0.91)
- Lite-Mono: A Lightweight CNN and Transformer Architecture for Self-Supervised Monocular Depth Estimation (CVPR 2023 · similarity 0.90)
- Self-Supervised Monocular Trained Depth Estimation Using Self-Attention and Discrete Disparity Volume (CVPR 2020 · similarity 0.90)
- The Temporal Opportunist: Self-Supervised Multi-Frame Monocular Depth (CVPR 2021 · similarity 0.90)
- Improved Monocular Depth Prediction Using Distance Transform Over Pre-semantic Contours with Self-supervised Neural Networks (CVPR 2025 · similarity 0.89)
- DME: Unveiling the Bias for Better Generalized Monocular Depth Estimation (AAAI 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.