Shadow removal is a challenging task as it usually requires the detection or annotation of shadows as well as semantic understanding of the scene. In this chapter, we introduce a shadow removal algorithm based on the shadow linear model and then introduce an end-to-end shadow removal algorithm based on deep learning without relying on shadow detection. Our deep neural network (DeshadowNet) is designed with a multi-context architecture, where the output shadow matte is predicted by embedding information from three different perspectives, i.e., high-level semantic context, mid-level appearance information, and local image details. Experimental results show that DeshadowNet is adaptive to shadows with widely varying penumbra widths and works well for shadows cast on different semantic regions. For highlight removal, in this chapter, we derive a global color-lines constraint from dichromatic reflection model to robustly recover specular and diffuse reflection. Our key observation is that each image pixel lies along a color line in normalized RGB space and the different color lines representing distinct diffuse chromaticities intersect at one point, namely the illumination chromaticity. For pixels along the same color line, they spread over the entire image and their distances to the illumination chromaticity reflect the amount of specular reflection components. With global (non-local) information from these color lines, our method can effectively separate specular and diffuse reflection components in a pixel-wise manner for a single image, and it is suitable for real-time applications. Experimental results on synthetic and real images show that our method performs better than the existing methods in terms of both specular reflection removal accuracy and computational efficiency.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Shadow and Highlight Removal


    Weitere Titelangaben:

    Research on Intelligent Manufacturing


    Beteiligte:
    Tian, Jiandong (Autor:in)


    Erscheinungsdatum :

    2021-12-02


    Format / Umfang :

    33 pages





    Medientyp :

    Aufsatz/Kapitel (Buch)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch




    Error diffusion with homogeneous highlight and shadow response [2949-38]

    Eschbach, R. / SPIE | British Library Conference Proceedings | 1996



    Highlight removal by illumination-constrained inpainting

    PingTan, / Lin, / Long Quan, et al. | IEEE | 2003


    Highlight Removal by Illumination-Constrained Inpainting

    Tan, P. / Lin, S. / Quan, L. et al. | British Library Conference Proceedings | 2003