U$^{2}$Flow: Uncertainty-Aware Unsupervised Optical Flow Estimation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sun, Xunpei, Lin, Wenwei, Chang, Yi, Chen, Gang
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917400398528512
author Sun, Xunpei
Lin, Wenwei
Chang, Yi
Chen, Gang
author_facet Sun, Xunpei
Lin, Wenwei
Chang, Yi
Chen, Gang
contents Unsupervised optical flow methods typically lack reliable uncertainty estimation, limiting their robustness and interpretability. We propose U$^{2}$Flow, the first recurrent unsupervised framework that jointly estimates optical flow and per-pixel uncertainty. The core innovation is a decoupled learning strategy that derives uncertainty supervision from augmentation consistency via a Laplace-based maximum likelihood objective, enabling stable training without ground truth. The predicted uncertainty is further integrated into the network to guide adaptive flow refinement and dynamically modulate the regional smoothness loss. Furthermore, we introduce an uncertainty-guided bidirectional flow fusion mechanism that enhances robustness in challenging regions. Extensive experiments on KITTI and Sintel demonstrate that U$^{2}$Flow achieves state-of-the-art performance among unsupervised methods while producing highly reliable uncertainty maps, validating the effectiveness of our joint estimation paradigm. The code is available at https://github.com/sunzunyi/U2FLOW.
format Preprint
id arxiv_https___arxiv_org_abs_2604_10056
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle U$^{2}$Flow: Uncertainty-Aware Unsupervised Optical Flow Estimation
Sun, Xunpei
Lin, Wenwei
Chang, Yi
Chen, Gang
Computer Vision and Pattern Recognition
Unsupervised optical flow methods typically lack reliable uncertainty estimation, limiting their robustness and interpretability. We propose U$^{2}$Flow, the first recurrent unsupervised framework that jointly estimates optical flow and per-pixel uncertainty. The core innovation is a decoupled learning strategy that derives uncertainty supervision from augmentation consistency via a Laplace-based maximum likelihood objective, enabling stable training without ground truth. The predicted uncertainty is further integrated into the network to guide adaptive flow refinement and dynamically modulate the regional smoothness loss. Furthermore, we introduce an uncertainty-guided bidirectional flow fusion mechanism that enhances robustness in challenging regions. Extensive experiments on KITTI and Sintel demonstrate that U$^{2}$Flow achieves state-of-the-art performance among unsupervised methods while producing highly reliable uncertainty maps, validating the effectiveness of our joint estimation paradigm. The code is available at https://github.com/sunzunyi/U2FLOW.
title U$^{2}$Flow: Uncertainty-Aware Unsupervised Optical Flow Estimation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.10056