Revision history [back]

I'm an opencv newbie wanting to explore computer vision. I was just starting a pet project of a rock paper scissors game where you can take a picture of an actual rock, paper or scissors for selection. I was using this page http://blog.yhathq.com/posts/image-classification-in-Python.html as a tutorial for image classification, and I figured that I could make it more accurate by introducing an edge detection step to simplify the image before processing (bc rock, paper and scissors all have different profiles [blob, rectangle, v shape]). I thought it may be a good idea to get some second opinions before starting because I've never done anything like this before and wanted to know if my game plan and the tutorial I'm following are sound, and what other peoples approaches may be.