Based on the complementary vision sensor (CVS), Tianmouc, we propose $\textbf{S}$patio-$\textbf{T}$emporal Difference $\textbf{G}$uided $\textbf{D}$eblur $\textbf{N}$et (STGDNet) for motion deblurring. It achieves strong performance in real-world extreme blur scenarios.
Drag the slider to compare the blurred input with our deblurred result. The real-world dataset is available here.
Given a single blurred frame as input, our method reconstructs the motion within the exposure time and generates a clear video.
@InProceedings{Meng_2026_CVPR,
author = {Meng, Yapeng and Yang, Lin and Chen, Yuguo and Chen, Xiangru and Wang, Taoyi and Wang, Lijian and Yang, Zheyu and Lin, Yihan and Zhao, Rong},
title = {Spatio-Temporal Difference Guided Motion Deblurring with the Complementary Vision Sensor},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2026},
pages = {37496-37506}
}Project page template is borrowed from DreamFusion.