Research map / Video / Video segmentation and tracking
Learning Object Depth from Camera Motion and Video Object Segmentation
Brent A. Griffin, Jason J. Corso
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 learned ODN_l instead of VOS-DE least squares (n=10) | object depth estimation · ODMS All Sets | mean percent error: 27.5 -> 19.0 | Improved |
| use ODN_d (depth-based prior) instead of ODN_l on Driving | ODN · ODMS Driving | mean percent error: 30.1 -> 47.5 | Hurt |
| use ODN_d instead of ODN_l on Simulated Normal | ODN · ODMS Simulated Normal | mean percent error: 8.3 -> 5.1 | Improved |
| reduce number of observations from 10 to 2 for ODN_l (ablation) | ODN_l · ODMS All Sets | mean percent error: 19.0 -> 20.4 | Hurt |
| reduce number of observations from 10 to 2 for VOS-DE (ablation) | VOS-DE · ODMS All Sets | mean percent error: 27.5 -> 50.3 | Hurt |
| reduce number of observations from 10 to 2 for ODN_dbar (ablation) | ODN_dbar · ODMS All Sets | mean percent error: 19.0 -> 22.7 | Hurt |
| reduce number of observations from 10 to 2 for ODN_d (ablation) | ODN_d · ODMS All Sets | mean percent error: 20.5 -> 21.6 | Hurt |
| add perturbation training data to ODN_l (ablation) | ODN_l · ODMS Simulated Perturb | mean percent error: 18.2 -> 13.0 | Improved |
| add perturbation training data to ODN_d (ablation) | ODN_d · ODMS Simulated Perturb | mean percent error: 11.2 -> 8.2 | Improved |
| add perturbation training data to ODN_d (ablation) | ODN_d · ODMS Driving | mean percent error: 47.5 -> 60.9 | Hurt |
| add perturbation training data to ODN_l (ablation) | ODN_l · ODMS All Sets | mean percent error: 19.0 -> 18.8 | Improved |
| add radial input image to ODN_l (ablation) | ODN_l · ODMS All Sets | mean percent error: 19.0 -> 17.8 | Improved |
Closest papers by meaning
- Depth From Camera Motion and Object Detection (CVPR 2021 · similarity 0.90)
- Depth Prediction without the Sensors: Leveraging Structure for Unsupervised Learning from Monocular Videos (AAAI 2019 · similarity 0.88)
- Depth From Videos in the Wild: Unsupervised Monocular Depth Learning From Unknown Cameras (ICCV 2019 · similarity 0.88)
- A Survey on Deep Learning Technique for Video Segmentation (TPAMI 2022 · similarity 0.88)
- Learning to Segment Moving Objects (IJCV 2018 · similarity 0.88)
- Multi-Object Discovery by Low-Dimensional Object Motion (ICCV 2023 · similarity 0.88)
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.
