Technical Briefs

    Image Processing

    Thursday, 21 November

    16:15 - 18:00

    Room S223

    Dense Scattering Layer Removal

    We propose a new model, together with advanced optimization, to separate a thick scattering media layer from a single natural image. Our method addresses the critical issue; originally unnoticeable impurities will be greatly magnified after removing the scattering media layer; with transmission-aware optimization.


    Qiong Yan, The Chinese University of Hong Kong
    Li Xu, The Chinese University of Hong Kong
    Jiaya Jia, The Chinese University of Hong Kong

    Decolorization: Is Rgb2gray() Out?

    Investigation in the RGB2GRAY model should be enough to solve the decolorization problem.


    Yibing Song,City University of Hong Kong
    Linchao Bao,City University of Hong Kong
    Xiaobin Xu,City University of Hong Kong
    Qingxiong Yang,City University of Hong Kong

    Dense Isocontour Imaging

    We present an imaging technique intended to explore multiscale image structures, represented by isophotes, based on a view dependent periodic transfer function with the period depending on the gradient magnitude of the underlying scalar function such as to create a dense visualization independent of the gradient magnitude.


    Victor Matvienko, Saarland University, Intel Visual Computing Institute
    Jens Krueger, Ivda, Intel Visual Computing Institute, University of Duisburg-Essen.

    Image-Pair-Based Anisotropic Material Modeling

    A method for modeling anisotropic materials with only two images taken under different lighting conditions by an ordinary camera. These images are used to obtain BRDF parameters, orientation field and height field for the material. With these information, new images can be rendered under new lighting and viewing conditions.


    Jie Feng, Peking University
    Wangyu Xiao, Peking University
    Bingfeng Zhou, Peking University

    Fast Multi-Scale Detail Decomposition via Accelerated Iterative Shrinkage

    This paper presents a fast multi-scale detail decomposition method. Smoothing is accelerated using proximal operators with a first-order approximation by setting new shrinkage operators and an efficient warm-start solution. The new method permits to perform faster multi-scale detail decomposition while still producing a suitable photographic look.


    Hicham Badri, INRIA Bordeaux Sud-Ouest
    Hussein Yahia, INRIA Bordeaux Sud-Ouest
    Driss Aboutajdine, Mohammed V-Agdal University