“Real-time stereo vision using semi-global matching on programmable graphics hardware” by Rosenberg, Davidson, Muller and Han

  • ©Ilya D. Rosenberg, Philip L. Davidson, Casey M. R. Muller, and Jefferson Y. Han




    Real-time stereo vision using semi-global matching on programmable graphics hardware



    Extracting depth information from pairs of stereo images in real-time allows interactive 3d operations on live video streams. We present a novel adaptation of the Semi-Global Matching (SGM) Algorithm [Hirschmüller 2005] to graphics processing (GPU) hardware, which operates in real-time and frees the CPU for other tasks. We also present shaders which use the extracted depth information to manipulate the video stream in real time.


    1. Hirschmüller, H. 2005. Accurate and Efficient Stereo Processing by Semi-Global Matching and Mutual Information. In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, vol. 2, 807–814.
    2. Scharstein, D. and Szeliski, R. 2002. A Taxonomy and Evaluation of Dense Two-Frame Stereo Correspondence Algorithms. In International Journal of Computer Vision 47, 1–3 (Apr. 2002), 7–42.

ACM Digital Library Publication:

Overview Page: