Applet Associate Points
After interest points have been detected and described, the next step when computing the transformation between two images is to associate the features in one image to the other. Association is when two features detected in different images are declared as being two views of the same physical object. How well two objects are associated depends on the technique used to associate them and on the descriptive power of the feature descriptor.
The applet below displays associated features between two images using colored lines. Click on a feature to only see that feature being displayed in each image. Use the toolbar on top to change the input image, interest point detection algorithm, region descriptor, and association algorithm.
What to look for:
- Which descriptors perform best in different images.
- How does the camera orientation effect association?
- How well does forwards-backwards validation prune more false positives compared to greedy?
- Which interest point detectors work best with pixel value (Pixel and NCC) based descriptors?
Related Tutorials/Example Code:
alt="Your browser understands the <APPLET> tag but is not running the applet, for some reason."
Your browser is completely ignoring the <APPLET> tag!