The digital future

Tathagata Chakraborty

Computer Vision - A Modern Approach David A. Forsyth and Jean Ponce Pearson Education

The subject of ‘Computer Vision’ has been around for some decades now; and the literature in this field is vast in volume and diverse in scope and application. In fact, this is the reason why it is difficult, if not impossible, to find a general book on computer vision. In addressing this need for a comprehensive as well as a ready reference, this book definitely does a commendable job. Almost all topics starting from generation of images to their interpretation, understanding and retrieval are comprehensively covered. Ideally meant as a reference for professionals working in related fields and application areas, this book also provides a one-stop solution to the undergraduate (in computer science or electrical engineering) struggling with tons of unsorted notes. However, the book might not be suitable for a seasoned professional, as he/she may find the topics of interest not covered to the required depth. Moreover, active professionals might not find the book complete. Although the authors have tried their best to this extent, a complete solution would have doubled the already bloated size of the book (handbooks on ‘Image Processing’ proper and ‘Pattern Recognition’ are recommended and might bridge the gap in such cases). In trying to address a broader audience though, the book loses some coherence in the contents within a chapter as also the general flow between them.

Traditionally, in our country the broader subject of computer vision is neglected in preference to specific topics of digital image/signal processing in most graduate and undergraduate courses. This is partly due to the expensive equipment required to set up a laboratory and partly due to the lack of good teaching resour ces—the latter should get addressed with the publication of this low-priced Indian edition. A good set of interesting exercises have been included at the end of each chapter, and a solutions manual is also available. The ‘Programming Assign ments’ at the end are worth implementing and would provide the beginner with a chance to perceptually understand the complex mathematical processes of computer vision. The book also contains a sufficiently large number of images that ease the process of understanding. However, in the low priced (paperback) edition of this book, a large number of the images are not even interpretable, let alone be made sense of by the techniques of computer vision. Hence I would advise a keen and prospective buyer to go for the hardbound edition (that would make you poorer by around $80-$90, plus shipping).

The author is a software engineer with Geometric Software