exFAST and High-Performance Sparse Stereo Matcher (HPSSM)

The provided source code package contains an efficient sparse stereo analysis algorithm that applies a dense consistency check, leading to accurate matching results. Matching accuracy is further improved by the new feature detector exFAST (extended FAST), which is based on FAST but exhibits a less clustered feature distribution. This feature detector leads to a superior performance of our stereo analysis algorithm. For details about both methods please consult our publication [1].

If you use our code for your own research, we kindly ask that you cite our paper [1] in any related publications.
exFAST FAST Sparse Stereo Matching


The source code is published under the GNU General Public License version 3, of which a copy is distributed with the code. Exempt from this license is the FAST9 detector (file: src/sparsestereo/fast9-inl.h), which was originally written by Edward Rosten and has only been modified by us.

The source code makes heavy use of the GCC vector extensions to take advantage of the SSE instruction sets. Hence, the code cannot be complied with a different compiler without making major modifications.


exFAST_SparseStereo.tar.bz2 (361 KB)


Konstantin Schauwecker, konstantin.schauwecker-NO_SPAM-@uni-tuebingen.de


[1] Konstantin Schauwecker, Reinhard Klette, and Andreas Zell. A New Feature Detector and Stereo Matching Method for Accurate High-Performance Sparse Stereo Matching. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 5171-5176, Vilamoura, Algarve, Portugal, October 2012. IEEE. [ link | pdf ]