Local photometric features have become popular as a practical and effective approach to image matching and recognition. They are distinctive as well as robust to occlusion and clutter. Recently these features have been extended to be scale and affine invariance which allows matching and recognition in the presence of large scale and viewpoint changes. In this short course we present these features and describe their application to image matching, 3D recognition and robot navigation as well as texture classification and detection of object categories. We will demo a system that uses these features to perform object recognition in real time.
8:00- 8:30   Introduction (Schmid), introduction.ppt
8:30- 9:00   Scale invariant interest points and SIFT descriptors (Lowe), scale_sift.ppt
9:00- 9:25   Affine invariance of interest points and applications (Schmid), affine.ppt
9:25- 9:45   Evaluation of interest points and descriptors (Schmid), evaluation.ppt
10:15-11:15   Object recognition system, demo, applications (Lowe), recog_system.ppt
11:15-11:45   Recognition of textures and object classes (Schmid), recognition.ppt
11:45-12:00   Future directions and discussion (Lowe), future.ppt
Related publications (more publications):
Local Greyvalue Invariants for Image Retrieval. C. Schmid and R. Mohr. In Pattern Analysis and Machine Intelligence, 1997.
Evaluation of Interest Point Detectors. C. Schmid, R. Mohr and C. Bauckhage. In International Journal of Computer Vision, 37(2), 151-172, 2000.
Indexing based on scale invariant interest points. K.Mikolajczyk and C.Schmid. In International Conference on Computer Vision, 525-531, 2001.
An affine invariant interest point detector. K.Mikolajczyk and C.Schmid. In European Conference on Computer Vision, vol. 1, 128--142, 2002.
A performance evaluation of local descriptors. K. Mikolajczk and C. Schmid. In IEEE Conference on Computer Vision and Pattern Recognition, 2003. pdf
F. Rothganger, S. Lazebnik, C. Schmid and J. Ponce. In IEEE Conference on Computer Vision and Pattern Recognition, 2003. pdf
Sparse texture representation using affine-invariant neighborhoods. S. Lazebnik, C. Schmid and J. Ponce. In IEEE Conference on Computer Vision and Pattern Recognition, 2003. pdf
Code : link to come