paper-with-me

홈 › Papers

WAFT: Warping-Alone Field Transforms for Optical Flow

2025-06-26 · Yihan Wang, Jia Deng

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 and KITTI benchmarks, achieves the best zero-shot generalization on KITTI, while being up to 4.1x faster than methods with similar performance. Code and model weights are available at https://github.com/princeton-vl/WAFT.

📄 PDF Abstract BibTeX arXiv:2506.21526

Code (1)

princeton-vl/waft 공식 구현 pytorch

Tasks

Optical Flow EstimationZero-shot Generalization

Similar Papers 제목 키워드 기반

WAFT-Stereo: Warping-Alone Field Transforms for Stereo Matching

2026-03-25 · Yihan Wang, Jia Deng arxiv

We introduce WAFT-Stereo, a simple and effective warping-based method for stereo matching. WAFT-Stereo demonstrates that cost volumes, a common design used in many leading methods, are not necessary for strong performanc…

Forecasting Future Instance Segmentation with Learned Optical Flow and Warping

2022-11-15 · Andrea Ciamarra, Federico Becattini, Lorenzo Seidenari, Alberto del Bimbo

For an autonomous vehicle it is essential to observe the ongoing dynamics of a scene and consequently predict imminent future scenarios to ensure safety to itself and others. This can be done using different sensors and …

Instance SegmentationOptical Flow EstimationSemantic Segmentation

DS-PASS: Detail-Sensitive Panoramic Annular Semantic Segmentation through SwaftNet for Surrounding Sensing

2019-09-17 · Kailun Yang, Xinxin Hu, Hao Chen, Kaite Xiang 외

Semantically interpreting the traffic scene is crucial for autonomous transportation and robotics systems. However, state-of-the-art semantic segmentation pipelines are dominantly designed to work with pinhole cameras an…

DecoderSegmentationSemantic SegmentationVisual Odometry

LookingGlass: Generative Anamorphoses via Laplacian Pyramid Warping

2025-01-01 · CVPR 2025 1 · Pascal Chang, Sergio Sancho, Jingwei Tang, Markus Gross 외

Anamorphosis refers to a category of images that are intentionally distorted, making them unrecognizable when viewed directly. Their true form only reveals itself when seen from a specific viewpoint, which can be thr…

valid

ProWAFT: A ROMA-LPD Instance for Workload-Aware and Dynamic Fault Tolerance in FPGA-Based CNN Accelerators

2026-07-02 · Xinxin Chen, Haoran Qiao, Yiming Guo, Kecheng Luo 외 arxiv

SRAM-based FPGAs provide an attractive platform for energy- and latency-constrained CNN inference at the network edge, yet transient faults can lead to silent errors that compromise reliability. Always-on redundancy (e.g…