Compact Signatures for High-Speed Interest Point Description and Matching

Prominent feature point descriptors such as SIFT and SURF allow reliable real-time matching but at a compu- tational cost that limits the number of points that can be handled on PCs, and even more on less powerful mobile devices. A recently proposed technique that relies on statis- tical classification to compute signatures has the potential to be much faster but at the cost of using very large amounts of memory, which makes it impractical for implementation on low-memory devices. In this paper, we show that we can exploit the sparseness of these signatures to compact them, speed up the compu- tation, and drastically reduce memory usage. We base our approach on Compressive Sensing theory. We also high- light its effectiveness by incorporating it into two very dif- ferent SLAM packages and demonstrating substantial per- formance increases.

• Type: Conference
• Year: 2009
• Where: ICCV
• Authors: Michael Calonder, Vincent Lepetit, Pascal Fua, Kurt Konolige, James Bowman, Patrick Mihelich
• Keywords: Compact Signatures, Computer Vision, High-Speed Interest Point Description, ICCV 2009