tags:

views:

536

answers:

1

Hi

I'm starting out studying some computer vision (on what I believe would be called "grad school level" in the US) and wish to learn OpenCV. The book "Learning OpenCV" seems like a good place to start, but as far as I can see after playing around, it seems that OpenCV has undergone some quite major changes in version 2.0+.

Is the book still relevant?

+1  A: 

Haha, I asked myself the same question after buying it. Apart from being a manual for the library, it also briefly describes several aspects of computer vision, so it's useful as a handbook of sorts as well. Also, since most of the code is in terms of IplImage, it may seem useless but you can always replace it with the Mat structure with a little effort. Also, some of the concepts are explained well (like the quad-edge structure in Delaunay triangulation), so that's very useful.

In short, it serves as a

  1. Handbook of computer vision
  2. Example code source
  3. Concepts behind the code such as data structures, etc. are explained
Jacob
I second that. The authors use OpenCV 1.1pre, and there have been many changes to OpenCV since then. That book needs an update for sure. Better to look in the darker corners of the internet for a pdf version, because you will not use it very much.
typoknig
Also, there's no harm in just using OpenCV 1.1. I bought the book and learned it along with 1.1. You will find that most of the non-official documentation on the web is still for for OpenCV 1 and so unless there is some compelling reason to use 2.0, I would just stick with that for now.
AndyL
If your question is whether there is a compelling reason to use 2.0, perhaps you should give it a try. Overloaded operator + automatic memory management is awesome (if used right).
Dat Chu