Proceedings: GI 2010

Real-time video matting using multichannel poisson equations

Minglun Gong, Liang Wang, Ruigang Yang, Yee-Hong Yang

Proceedings of Graphics Interface 2010: Ottawa, Ontario, Canada, 31 May - 2 June 2010, 89-96

  • BibTex

    @inproceedings{Gong:2010:,
    author = {Gong, Minglun and Wang, Liang and Yang, Ruigang and Yang, Yee-Hong},
    title = {Real-time video matting using multichannel poisson equations},
    booktitle = {Proceedings of Graphics Interface 2010},
    series = {GI 2010},
    year = {2010},
    issn = {0713-5424},
    isbn = {978-1-56881-712-5},
    location = {Ottawa, Ontario, Canada},
    pages = {89--96},
    numpages = {8},
    publisher = {Canadian Human-Computer Communications Society},
    address = {Toronto, Ontario, Canada},
    }

Abstract

This paper presents a novel matting algorithm for processing video sequences in real-time and online. The algorithm is based on a set of novel Poisson equations that are derived for handling multichannel color vectors, as well as the depth information captured. A simple yet effective approach is also proposed to compute an initial alpha matte in the color space. Real-time processing speed is achieved through optimizing the algorithm for parallel processing on the GPUs. To process live video sequences online and autonomously, a modified background cut algorithm is implemented to separate foreground and background, the result of which guides the automatic trimap generation. Quantitative evaluation on still images shows that the alpha mattes extracted using the presented algorithm is much more accurate than the ones obtained using the global Poisson matting algorithm and are comparable to that of other state-of-the-art offline image matting techniques.