Blur robust optical flow using motion channel

NEUROCOMPUTING

卷:220

Article

Li, Wenbin^1,5  Chen, Yang²  Lee, JeeHang³  Ren, Gang⁴  Cosker, Darren⁵ 

[1] UCL, Dept Comp Sci, London, England

[2] Imperial Coll London, Hamlyn Ctr, London, England

[3] Univ Bath, Dept Comp Sci, Bath, Avon, England

[4] Xiamen Univ Technol, Sch Digital Art, Xiamen, Peoples R China

[5] Univ Bath, CAMERA, Bath, Avon, England

关键词: Optical flow;    Computer vision;    Image deblurring;    Directional filtering;    RGB-motion imaging;   

DOI  :  10.1016/j.neucom.2016.03.105

来源: Elsevier

【 摘 要 】

It is hard to estimate optical flow given a realworld video sequence with camera shake and other motion blur. In this paper, we first investigate the blur parameterisation for video footage using near linear motion elements. We then combine a commercial 3D pose sensor with an RGB camera, in order to film video footage of interest together with the camera motion. We illustrate that this additional camera motion/trajectory channel can be embedded into a hybrid framework by interleaving an iterative blind deconvolution and warping based optical flow scheme. Our method yields improved accuracy within three other state-of-the-art baselines given our proposed ground truth blurry sequences; and several other realworld sequences filmed by our imaging system.

【 授权许可】

Free