SURF
SURF ( English , Speeded Up Robust Features , freely translated: "Accelerated, robust features") is an algorithm by Herbert Bay et al. for fast and robust recognition of image features for machine vision . The application of this algorithm is patented. SURF replaces the Gaussian filters used in SIFT with mean value filters , which can be calculated using integral images with a constant expenditure of time.
Implementations
- The get_surf_points (image) function is contained in the Dlib program library .
- In the OpenCV program library under opencv_contrib / modules / xfeatures2d / non-free algorithms such as SURF and SIFT are available in the source code. This module is not part of the standard delivery.
Web links
- SURF on Github
- Homepage of the SURF algorithm
- Presentation: GPU Accelerating Speeded-Up Robust Features (PDF file; 4.22 MB)
- Alexandre Chariot, Renaud Keriven: GPU-boosted online image matching (PDF file; 666 kB)
Individual evidence
- ↑ US 2009238460. In: Espacenet - Bibliographical Data. September 24, 2009, accessed December 13, 2018 .
- ↑ Herbert Bay, Tinne Tuytelaars and Luc Van Gool: SURF: Speeded Up Robust Features ( Memento from March 19, 2015 in the Internet Archive ) (PDF; 723 kB) , Proceedings of the 9th European Conference on Computer Vision, Springer Verlag, 2006
- ↑ Herbert Bay, Andreas Ess, Tinne Tuytelaars, and Luc Van Gool: Speeded-Up Robust Features (SURF). (PDF) ETH Zurich, Katholieke Universiteit Leuven, September 10, 2008, accessed on December 13, 2018 (English).
- ↑ dlib C ++ Library - surf_ex.cpp. Retrieved January 8, 2019 .
- ↑ opencv_contrib: Repository for OpenCV's extra modules. In: GitHub . OpenCV, December 13, 2018, accessed December 13, 2018 .