Project 1.  Feature Detection and Feature Descriptor


Submit your source code (c/matlab) and your report.




1.1             Interest Point Detector


Implement one or more keypoint detectors and compare their performance (with your own or with your classmate’s detector) ;

----- Options of detector --------

a.   Harris corner

-          Laplacian or Difference of Gaussian

-          Forstner-Harris Hessian with different formula variants given in (4.9, 4.10, 4.11)



-          Implement non-maximal suppression, such as [Brown, Szeliski, and Winder 2005]

b.   SIFT

c. Any other detectors



-         Show the response map for each detector

-         Show the strongest 250 and 500 interest points with different suppression radius



1.2           Interest Point Descriptor


Implement one or more descriptors (steered to local scale and orientation) and compare their performance.

-          Contrast-normalized patches (Brown, Szeliski, and Winder 2005)

-          SIFT (Lowe 2004)

-          GLOH (Mikolajczyk and Schmid 2005)

-  Any other descriptors