I'm not sure were to start ... this would be ok as a beginners books from a instruction viewpoint and people that really wanted to use opencv2. I think he explained opencv2 clearly ... but from a "machine vision" viewpoint the temporal aspect is very important ... and that was at the end of the book(i.e. he is mainly using static images and not comparing images sensors in time ... in almost all his examples). From a actual library perspective ... this is a bunch of easy matrix routines, not network mesh based, so the library is more old school image processing(i.e. the title of saying computer vision ... is missleading ... since this has very little to do with modern vision algorithms). It's terribly out of date from a viewpoint on where the state of the art is for vision processing ... (like 20 years) ... maybe that's in part of the library not included in the book. Using python/numpy/pil/etc.. is way easier to teach vision theory ... so I'm not sure who this book is for.