Tensor Voting Download Page


We present a method for Perceptual Grouping in Computer Vision for the inference of structure from sparse, noisy data in 2-D and 3-D.
Over the past few years, we have developed a unified Tensor Based framework to formalize this problem. The method is not iterative, uses no hard thresholds, and there is a single free parameter. We have successfully demonstrated the approach on different early vision tasks and technical applications.
For details please refer to our book
cover A Computational Framework for Segmentation and Grouping
by Gérard Medioni, Mi-Suen Lee, and Chi-Keung Tang
Elsevier 2000
As a brief introduction we suggest to read the preface and the software tutorial.

Download Software

There are three free software packages (beta-version) available. The 2D and 3D tensor voting systems, and the ViewWorld program for visualization. ViewWorld accepts input files accepted, and outputs generated by the two systems.
  • 2D system
  • 3D system
  • ViewWorld

  • Please register first before you download the software.

    Further Information

  • Prof. Gérard G. Medioni, Mi-Suen Lee, Chi-Keung Tang

  • Institute for Robotics and Intelligent Systems, University of Southern California, E-mail: tensorvt@iris.usc.edu