Research map / Label-efficient and robust learning / Self-supervised and contrastive
Learning by Analogy: Reliable Supervision From Transformations for Unsupervised Optical Flow Estimation
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 |
|---|---|---|---|
| Use transformed predictions of original images to supervise transformed-image predictions (ablation) | ARFlow · Sintel Clean validation | AEPE: 2.53 -> 2.04 | Helped |
| Use transformed predictions of original images to supervise transformed-image predictions (ablation) | ARFlow · Sintel Final validation | AEPE: 3.47 -> 2.97 | Helped |
| Add spatial transformation (ablation) | ARFlow · Sintel Clean validation | AEPE: 2.53 -> 2.14 | Helped |
| Add appearance transformation (ablation) | ARFlow · Sintel Clean validation | AEPE: 2.53 -> 2.40 | Helped |
| Add occlusion transformation (ablation) | ARFlow · Sintel Clean validation | AEPE: 2.53 -> 2.39 | Helped |
| Apply heavy transformations directly as ordinary augmentation (ablation) | ARFlow · Sintel Clean validation | AEPE: 2.53 -> 2.71 | Hurt |
| Use augmentation distillation instead of the proposed single-stage consistency training (ablation) | ARFlow · Sintel Clean validation | AEPE: 2.53 -> 2.36 | Helped |
| Use augmentation regularization rather than augmentation distillation (ablation) | ARFlow · Sintel Clean validation | AEPE: 2.36 -> 2.04 | Helped |
| Reduce dense connections in the decoder while adding augmentation regularization (ablation) | PWC-Net · Sintel Clean validation | AEPE: 2.53 -> 2.04 | Helped |
| Share the decoder across pyramid levels while adding augmentation regularization (ablation) | PWC-Net · Sintel Clean validation | AEPE: 2.30 -> 1.95 | Helped |
| Add multiple frames to the shared-decoder model (ablation) | ARFlow · Sintel Clean validation | AEPE: 1.95 -> 1.89 | Helped |
| Add AutoAugment within the proposed framework (ablation) | ARFlow · Sintel Clean validation | AEPE: 2.53 -> 2.30 | Helped |
Closest papers by meaning
- Unsupervised Deep Learning for Optical Flow Estimation (AAAI 2017 · similarity 0.92)
- UnSAMFlow: Unsupervised Optical Flow Guided by Segment Anything Model (CVPR 2024 · similarity 0.90)
- Taming generative video models for zero-shot optical flow extraction (NeurIPS 2025 · similarity 0.88)
- OCAI: Improving Optical Flow Estimation by Occlusion and Consistency Aware Interpolation (CVPR 2024 · similarity 0.87)
- Recurrent Partial Kernel Network for Efficient Optical Flow Estimation (AAAI 2024 · similarity 0.87)
- ADFactory: An Effective Framework for Generalizing Optical Flow with NeRF (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.