Thomas Deselaers, Vittorio Ferrari

Overview

Welcome to this release of the global selfsimilarity [1]. From the methods in [1] this code allows to compute self-similarity hypercubes (SSHs). These SSHs can be obtained from a prototype assignment map M in two ways:
The first method extracts an SSH from a full image.
The second method extracts multiple SSHs from a single image jointly.

Downloads

References

1. Global and Efficient Self-Similarity for Object Classification and Detection
   Thomas Deselaers and Vittorio Ferrari,
   In Computer Vision and Pattern Recognition (CVPR), 2010.