It's rather problematic. I'd start with sciweavers ( http://www.sciweavers.org/ ), since they have a lot of readable, new papers, but the vast majority of the references are fairly math-heavy and somewhat removed from the applications. A very good resource is the Oreilly book ( http://www.amazon.com/Learning-OpenCV-Computer-Vision-Librar... ) on OpenCV, and the OpenCV source code ( https://code.ros.org/svn/opencv/trunk/opencv/ ). Imagemagick source ( http://www.imagemagick.org/script/index.php ) is also amazing for learning how basic operations are done in code.

Learning to read the style of older research papers and to pull applications out of them is very useful though. It's surprising how often something seemingly overacademized and useless, when translated into code, turns out to be very useful and smart.

O'Reilly "Learning OpenCV" is very well written and covers a wide range of topics.

http://www.amazon.com/Learning-OpenCV-Computer-Vision-Librar...