“Color Correction Algorithm based on Local Similarity of Stereo Images” by Lee and Lee

  • ©Yong-Ho Lee and In-Kwon Lee

  • ©Yong-Ho Lee and In-Kwon Lee

  • ©Yong-Ho Lee and In-Kwon Lee



Entry Number: 28


    Color Correction Algorithm based on Local Similarity of Stereo Images



    In stereoscopic 3D content creation utilizing stereo camera, luminance and color discrepancies between stereo images often exists. These discrepancies result in incorrect depth information extraction during post-production and cause visual fatigue for the audience. In stereoscopic color correction research, local methods are generally superior to global methods, because the stereo image has local color discrepancies. However, previous local methods [Wang et al. 2011] cannot manage specific local color discrepancies such as highlighting and various illumination conditions, because these methods only obtain and apply a sparse sampling of the correspondences on the image. Thus, they can generate biased color compensating results.
    The proposed method can solve the problems of the previous method by extracting relevant correspondence pair data for color transfers using a modified stereo matching algorithm, and by compensating the color using weighted sum of color differences considering local features that can represent local luminance and color discrepancies.


    1. Sun, X., Mei, X., Shaohui Jiao, Zhou, M., and Wang, H. 2011. Stereo matching with reliable disparity propagation. In 3D Imaging, Modeling, Processing, Visualization and Transmission (3DIMPVT), 2011 International Conference on, 132–139.
    2. Wang, Q., Yan, P., Yuan, Y., and Li, X. 2011. Robust color correction in stereo vision. In Image Processing (ICIP), 2011 18th IEEE International Conference on, 965–968.


Additional Images:

©Yong-Ho Lee and In-Kwon Lee ©Yong-Ho Lee and In-Kwon Lee


    This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (No.2011- 0028568).


ACM Digital Library Publication:

Overview Page: