SIFT algorithm based on openCV & GSL
While now I learn SIFT algorithm by David G.Lowe. and I study read his paper:Distinctive Image Features from Scale-Invariant Keypoints.
Reference:
1, http://web.engr.oregonstate.edu/~hess/index.html
2, http://www.opencv.org.cn/index.php/User:Lmyanglei
3, http://gnuwin32.sourceforge.net/packages/gsl.htm
4, http://www.gnu.org/software/gsl/
and 3,4 is for helping to understand “Gsl”
Talk about whole SIFT:
This algroithm can be used to perform reliable matching between different views of an object or scene. And their features are invarint to image scale and rotation,and are shown to provide robust matching across a substantial range of affine distortion, change in 3D viewpoint,addition of noise,and change in illumination.
And below is main features:
1, be invariant to image scaling and rotation, and partially invariant ro change in illumination and 3D camera viewpoint.
2, be well localized in both the spatial and frequency domains, reducing the probability of disruption by occlusion,clutter or noise.
3, be highly distinctive, and allow a single feature to be correctly matched with high probability against a large database of features, providing a basis for object and scene recognition
4, generate large numbers of features that densely cover the image over the full range of scales and locations
5, computation rapidly against large databases
And below is SIFT’s algorithm:
1, scale-space extrema detection
2, keypoint localization
3, orientation assignment
4, keypoint descriptor
And when I debugging program, I found there are two feactures : oxford-features &
lowe-features.
I download the resource and debugging…
If you want to learn SIFT, you’d better to my reference.