WAFT: Warping-Alone Field Transforms for Optical Flow
Yihan Wang, Jia Deng
Abstract
We introduce Warping-Alone Field Transforms (WAFT), a simple and effective method for optical flow. WAFT is similar to RAFT but replaces cost volume with high-resolution warping, achieving better accuracy with lower memory cost. This design challenges the conventional wisdom that constructing cost volumes is necessary for strong performance. WAFT is a simple and flexible meta-architecture with minimal inductive biases and reliance on custom designs. Compared with existing methods, WAFT ranks 1st on Spring, Sintel, and KITTI benchmarks, achieves the best zero-shot generalization on KITTI, while being 1.3 -4.1× faster than existing methods that have competitive accuracy (e.g., 1.3× than Flowformer++, 4.1× than CCMR+). Code and model weights are available at https://github.com/princeton-vl/WAFT .
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext f0a4a6fb-a3ec-4cbd-b608-fe969db490fbCited by top-tier papers5
- CoWTracker: Tracking by Warping instead of CorrelationZihang Lai, Eldar Insafutdinov, Edgar Sucar, Andrea VedaldiCVPR 2026 · 12 citations
- Efficient All-Pairs Correlation Volume Sampling for Optical Flow EstimationKarlis Martins Briedis, Markus Gross, Christopher SchroersCVPR 2026 · 1 citation
- RetimeGS: Continuous-Time Reconstruction of 4D Gaussian SplattingXuezhen Wang, Li Ma, Yulin Shen, Zeyu Wang et al.CVPR 2026
- Learning to Generate Highly Dynamic Videos using Synthetic Motion DataWonjoon Jin, Jiyun Won, Janghyeok Han, Qi Dai et al.CVPR 2026
- ARFlow: Auto-regressive Optical Flow Estimation for Arbitrary-Length Videos via Progressive Next-Frame ForecastingJiuming Liu, Mengmeng Liu, Siting Zhu, Yunpeng Zhang et al.ICLR 2026
Builds on29
- An Image is Worth 16x16 Words: Transformers for Image Recognition at ScaleAlexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn et al.ICLR 2021 · 21,477 citations
- Segment AnythingAlexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao et al.ICCV 2023 · 13,211 citations
- High-Resolution Image Synthesis with Latent Diffusion ModelsRobin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser et al.CVPR 2022 · 13,123 citations
- Vision Transformers for Dense PredictionRené Ranftl, Alexey Bochkovskiy, Vladlen KoltunICCV 2021 · 2,647 citations
- Depth Anything V2Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao et al.NeurIPS 2024 · 2,305 citations
Related papers
- GMFlow: Learning Optical Flow via Global MatchingHaofei Xu, Jing Zhang, Jianfei Cai, Hamid Rezatofighi et al.CVPR 2022 · 353 citations
- FlowSeek: Optical Flow Made Easier with Depth Foundation Models and Motion BasesMatteo Poggi, Fabio TosiICCV 2025 · 5 citations
- CRAFT: Cross-Attentional Flow Transformer for Robust Optical FlowXiuchao Sui, Shaohua Li, Xue Geng, Yan Wu et al.CVPR 2022 · 114 citations
- FlowFormer++: Masked Cost Volume Autoencoding for Pretraining Optical Flow EstimationXiaoyu Shi, Zhaoyang Huang, Dasong Li, Manyuan Zhang et al.CVPR 2023
- MEMFOF: High-Resolution Training for Memory-Efficient Multi-Frame Optical Flow EstimationVladislav Bargatin, Egor Chistov, Alexander Yakovenko, Dmitriy S. VatolinICCV 2025 · 11 citations
