ASIFT: An Algorithm for Fully Affine Invariant Comparison

• article
• demo
• archive

BibTeX info

Communicated by Guillermo Sapiro
Demo edited by Nicolas Limare

Abstract

If a physical object has a smooth or piecewise smooth boundary, its images obtained by cameras in varying positions undergo smooth apparent deformations. These deformations are locally well approximated by affine transforms of the image plane. In consequence the solid object recognition problem has often been led back to the computation of affine invariant image local features. The similarity invariance (invariance to translation, rotation, and zoom) is dealt with rigorously by the SIFT method The method illustrated and demonstrated in this work, Affine-SIFT (ASIFT), simulates a set of sample views of the initial images, obtainable by varying the two camera axis orientation parameters, namely the latitude and the longitude angles, which are not treated by the SIFT method. Then it applies the SIFT method itself to all images thus generated. Thus, ASIFT covers effectively all six parameters of the affine transform.

Download

• full text manuscript: PDF low-res. (2.7M) PDF (9M) ^[?]
• source code: TAR/GZ

Supplementary Materials

• Slides: one hour version, 20 minutes version
• Videos: magazine, facade
• Poster
• Dataset (tar.gz version), (zip version) : An image dataset for systematic evaluation of robustness to absolute and transition tilt of the image matching algorithms is available
• Executable version of the code (Windows) (version April 30, 2011).

Non-Reviewed Supplementary Materials

These files and information are provided by the authors and have not been reviewed.

• note: On the consistency of the SIFT method

History

• this article was converted to PDF on 2015-09-25
• the original version was published on 2011-02-24: manuscript
• Note from the editor: The source code was updated on April 30, 2011 to resolve a possible copyright issue. Only the SVD functions were replaced.
• Note from the editor: The source code was updated on December 13, 2015 to fix a bug reported by Xiaoyu Sun (Sichuan university), whom we thank heartily. The modification of the code affects lines 75 to 77 of file compute_asift_matches.cpp (function unique_match1).
• Note from the editor: The source code was updated on October 30, 2018 to avoid an undue exponentially growing number of threads. The modification of the code affects file compute_asift_matches.cpp. The previous version of the code is available here.
• Note from the editor: the manuscript of the article was modified on 2022-01-01 to include information about its editors. The original version of the manuscript is available here.
• Note from the editor: The source code was updated on January 26, 2022 to add a missing C++ header string.h in io_png.c; also some misleading indentations were corrected. The previous version of the code is available here.