Evaluation of local detectors on non-planar scenes

This paper presents for the first time a method to evaluate the performance of local detectors under viewpoint changes on complex, realistic, and practically relevant scenes. The main contribution is a method which allows the automatic verification of detected corresponding points and regions for non-planar scenes. Using this method the performances of 10 different local detectors were evaluated in a large scale experiment. A ranking of the different detectors has been established based on this evaluation.

[1]  Luc Van Gool,et al.  Content-Based Image Retrieval Based on Local Affinely Invariant Regions , 1999, VISUAL.

[2]  David G. Lowe,et al.  Object recognition from local scale-invariant features , 1999, Proceedings of the Seventh IEEE International Conference on Computer Vision.

[3]  Luc Van Gool,et al.  Wide Baseline Stereo Matching based on Local, Affinely Invariant Regions , 2000, BMVC.

[4]  Bernhard P. Wrobel,et al.  Multiple View Geometry in Computer Vision , 2001 .

[5]  Jiri Matas,et al.  Robust wide-baseline stereo from maximally stable extremal regions , 2004, Image Vis. Comput..

[6]  Gustavo Carneiro,et al.  Phase-Based Local Features , 2002, ECCV.

[7]  Cordelia Schmid,et al.  Evaluation of Interest Point Detectors , 2000, International Journal of Computer Vision.

[8]  Cordelia Schmid,et al.  Comparison of affine-invariant local detectors and descriptors , 2004, 2004 12th European Signal Processing Conference.

[9]  Andrew Zisserman,et al.  An Affine Invariant Salient Region Detector , 2004, ECCV.

[10]  Cordelia Schmid,et al.  A performance evaluation of local descriptors , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.