For Automatic Image Annotation, i try to use through haar training to train system for detecting hill,sky,ground,building,water and so on in order to annotate an query image automatically.
I don't think that cascade classifier will help you. AFAIK your task is called Scene Understanding, but I don't know what to recommend to you. I've seen this project, and results are looking promising, but it seems they're using 3D data. Here is another project: http://www.cs.uiuc.edu/homes/dhoiem/projects/context/index.html.
And here some more recent work on this topic: http://www.cs.unc.edu/~jtighe/Papers/CVPR13/index.html
Asked: 2013-07-30 11:49:01 -0600
Seen: 251 times
Last updated: Jul 31 '13
