Research map / Architectures and efficiency / RNNs and normalization
A Lightweight Optical Flow CNN —Revisiting Data Fidelity and Regularization
Tak-Wai Hui, Xiaoou Tang, Chen Change Loy
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 extra training loss, lr 6e-5, and LiteFlowNet2 changes instead of LiteFlowNet (Chairs) | LiteFlowNet · KITTI 2012 (Chairs-trained) | AEE: 4.73 -> 4.11 | Improved |
| raise learning rate from 5e-5 to 6e-5 (ablation) | LiteFlowNet2 · KITTI 2015 (Chairs-trained) | AEE: 12.01 -> 11.76 | Improved |
| add extra training loss at upsampled final flow (ablation) | LiteFlowNet2 · KITTI 2015 (Chairs-trained) | AEE: 11.76 -> 11.31 | Improved |
| add pyramid level 2 to NetE (level 3 -> level 2) (ablation) | LiteFlowNet · Sintel Clean (train) | AEE: 2.65 -> 2.48 | No clear effect |
| add pyramid level 2 to NetE (ablation) | LiteFlowNet · Sintel runtime | runtime (ms): 13.06 -> 52.51 | Hurt |
| add pseudo flow inference and regularization at level 2 (ablation) | LiteFlowNet2 · Sintel runtime | runtime (ms): 52.51ms to 8.95ms | Improved |
| add pseudo network at level 2 (ablation) | LiteFlowNet2 · Sintel testing set | AEE | No clear effect |
| add pseudo network at level 2 (ablation) | LiteFlowNet2 · KITTI testing set | AEE | Improved |
| fine-tune on Sintel+KITTI mixture instead of Sintel only (ablation) | LiteFlowNet2-ft · Sintel Clean (test) | AEE: 3.83 -> 3.48 | Improved |
| fine-tune on Sintel+KITTI mixture instead of Sintel only (ablation) | LiteFlowNet2-ft · Sintel Final (test) | AEE: 5.06 -> 4.69 | Improved |
| train LiteFlowNet with new fine-tuning protocol (ablation) | LiteFlowNet-ft · Sintel Clean (test) | AEE: 4.54 -> 4.01 | Improved |
| train LiteFlowNet with new fine-tuning protocol (ablation) | LiteFlowNet-ft · Sintel Final (test) | AEE: 5.38 -> 5.21 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- LiteFlowNet — LiteFlowNet: A Lightweight Convolutional Neural Network for Optical Flow Estimation (CVPR 2018) · better on 4 of 4 shared results
- PWC-Net — PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume (CVPR 2018) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- LiteFlowNet: A Lightweight Convolutional Neural Network for Optical Flow Estimation (CVPR 2018 · similarity 0.97)
- LiteFlowNet3: Resolving Correspondence Ambiguity for More Accurate Optical Flow Estimation (ECCV 2020 · similarity 0.93)
- Models Matter, So Does Training: An Empirical Study of CNNs for Optical Flow Estimation (TPAMI 2019 · similarity 0.93)
- Recurrent Partial Kernel Network for Efficient Optical Flow Estimation (AAAI 2024 · similarity 0.89)
- DPFlow: Adaptive Optical Flow Estimation with a Dual-Pyramid Framework (CVPR 2025 · similarity 0.87)
- FlowDiffuser: Advancing Optical Flow Estimation with Diffusion Models (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.
