“Patching of moving objects for ghosting-free HDR synthesis” by Zheng, Li, Zhu, Wu and Rahardja

  • ©Jinghong Zheng, Zhengguo Li, Zijian Zhu, Shiqian Wu, and Susanto Rahardja

  • ©Jinghong Zheng, Zhengguo Li, Zijian Zhu, Shiqian Wu, and Susanto Rahardja

Conference:


Type(s):


Title:

    Patching of moving objects for ghosting-free HDR synthesis

Presenter(s)/Author(s):



Abstract:


    It is very challenging to synthesize a ghosting-free high dynamic range (HDR) image using differently exposed low dynamic range (LDR) images with moving objects. To achieve this, an anti-ghosting scheme is required to detect and patch occluded motion regions. In this poster, a new optimization problem is formulated on patching motion regions along a differently exposed image sequence, which optimizes consistencies in both spatial and temporal domain. To solve this optimization problem, a hybrid patching scheme including pixel-level intensity mapping function and block-based template matching is proposed.

References:


    1. Li, Z. G., Rahardja, S., Zhu, Z. J., Xie, S. L., and Wu, S. Q. 2010. Movement detection for the synthesis of high dynamic range images. In IEEE Int. Conf. Image Processing, 3133–3136.
    2. Perez, P., Gangnet, M., and Blake, A. 2003. Poisson image editing. In ACM SIGGRAPH.
    3. Jacobs, K., Ward, G., and Loscos, C. 2005. Automatic HDRI generation of dynamic environments. In ACM SIGGRAPH.


Additional Images:

©Jinghong Zheng, Zhengguo Li, Zijian Zhu, Shiqian Wu, and Susanto Rahardja ©Jinghong Zheng, Zhengguo Li, Zijian Zhu, Shiqian Wu, and Susanto Rahardja ©Jinghong Zheng, Zhengguo Li, Zijian Zhu, Shiqian Wu, and Susanto Rahardja

ACM Digital Library Publication:



Overview Page: