Real-time 3D scene reconstruction

Various future video communication and media production applications require real-time processing and low delay on top of it. Our group has built up long lasting experience on

  • Porting of algorithms to graphics processors (GPU),
  • Re-design and implementation of algorithms towards parallelization,
  • Development of a 3D video processing software framework that exploits multi-threading and distributed processing on multiple PCs, and
  • Implementation of fast inter-PC and inter process communication using TCP/IP or Infiniband.

Selected publications

  1. W. Waizenegger, I. Feldmann, O. Schreer, P. Eisert:
    Scene Flow Constrained Multi-Prior Patch-Sweeping for Real-Time Upper Body 3D Reconstruction, Proceedings IEEE International Conference on Image Processing (ICIP), Melbourne, Australia, September 2013.
  2. W. Waizenegger, I. Feldmann, P. Eisert:
    Depth Driven Photometric and Geometric Image Registration for Real-Time Stereo Systems, 16th International Workshop on Vision, Modeling and Visualization (VMV), Berlin, Germany, pp. 25-32, October 4-6, 2011.
  3. W. Waizenegger, N. Atzpadin, O. Schreer, I. Feldmann:
    Patch-Sweeping with Robust Prior for High Precision Depth Estimation in Real-Time Systems, Proceedings of 18th International Conference on Image Processing (ICIP 2011), Brussels, Belgium, September 11-14, 2011.
  4. W. Waizenegger, I. Feldmann, P. Eisert, P. Kauff:
    Parallel High Resolution Real-Time Visual Hull on GPU, Proceedings of International Conference on Image Processing (ICIP), Cairo, Egypt, November 2009.
  5. I. Feldmann, N. Atzapadin, O. Schreer, J.-C. Pujol-Acolado, J.-L. Landabaso, O. Divorra Escoda:
    Multi-View Depth Estimation Based on Visual-Hull Enhanced Hybrid Recursive Matching for 3D Video Conference Systems, Proceedings of 16th International Conference on Image Processing (ICIP 2009), Kairo, Egypt, November 2009.