Both video analytics and business intelligence are relatively new areas of research. They are also some of the most active areas. Video analytics is concerned with visual analysis (similar to image understanding). The application of such analysis is wide and varies greatly. Business intelligence is a prime application area of video analytics, in areas such as consumer-driven applications and security. This book looks at video analytics techniques and provides a connection between these techniques and business intelligence applications.
I was impressed by this book for two reasons: it is very well organized in terms of topics and the sequence of chapters; and it covers a very new, multidisciplinary, and active area of research, providing details that will benefit interested readers. It is uncommon to start a review by stating the positives about the book, but these two factors will make the review simpler.
The book is divided into four parts of two to four chapters each. The first part, “Computational Vision,” contains interesting chapters on video analysis that should be of interest to machine vision specialists. I enjoyed the chapter on object detection and tracking; it gives a good overview of the topic. The other two chapters--one on CCTV and one on example-based search--are also well written and very technical in their details.
Part 2, “Demographics,” should interest business intelligence researchers in general and those working on consumer behavior analysis in particular. The four chapters in this part form a bridge from computational vision (Part 1) to behavior analysis (Part 3). The first two chapters in this section, “Human Age Estimation and Sex Classification” and “People Counter,” seem to be a continuation of Part 1 and give examples of applications. However, the later two chapters in this part, on crowd and consumer behavior analysis, provide an excellent segue to Part 3. All four chapters are very technical and cover a wide range of application techniques.
Parts 3, “Behavior Analysis,” and 4, “Systems,” contain two chapters each. One may argue whether these chapters needed to be in separate parts. Part 3 is an excellent continuation of the chapters in Part 2. Part 4, however, is less impressive and seems to be somewhat repetitive. The first chapter in Part 4, “Video Analytics for Business Intelligence,” which is also incidentally the title of the book, may have served better as an introduction to the book rather than appearing near the end. The second chapter in this part, on people queue statistics estimation, seems repetitive.
Overall, this is an excellent reference for researchers and professionals interested in working in this overlapping area of video analytics and business intelligence.