****** - Verified Buyer
4.5
The intended audience for this book are research and development professionals in image processing. This is not a book for video hobbyists, but instead you should at least be at the advanced undergraduate level in electrical engineering or computer science to fully grasp the concepts presented. The book's goal is to show a survey of well-known deinterlacing algorithms and then go into details of how you can improve upon these algorithms. The book is divided into three major parts: basic technology, system optimization, and the future of interlace, and contains the following chapters:1. Introduction2. Overview of de-interlacing algorithms3. Motion estimation on interlaced video4. Improving motion-vector accuracy5. On the optimization of de-interlacing6. MPEG-2 coding efficiency of interlaced versus progressive video7. Display formatsChapter 2 appears to be an expansion of a paper that the authors wrote entitled "De-interlacing an overview". When I first read that paper I found it to be the clearest I had ever read on the subject, yet their book manages to improve it further. The emphasis is on the authors' own algorithm -adaptive recursive deinterlacing.Likewise, chapter 3 is an expansion of the authors' paper "Motion estimation on interlaced video". The authors examine and compare the performance of block matching algorithms, true motion estimation, global-based models, and object-based motion models.Chapter 4 examines accuracy of motion vectors, improving block-based motion estimation, and using interpolation to improve accuracy.Chapter 5 examines the optimization of interlacing and introduces the Majority-Selection de-interlacer, which combines the strengths of several individual de-interlacing algorithms into a single output signal. This material is from the authors' paper "Majority selection de-interlacing; An advanced motion-compensated spatio-temporal interpolation technique for interlaced video".Chapter 6 looks at the MPEG-2 video coding standard and discusses the efficiency of interlaced versus progressive video on an MPEG-2 digital channel. This chapter draws heavily from the authors' paper "On Coding Efficiency and Scan-Rate Conversion".Chapter 7 borrows heavily from the authors' paper "Towards an Optimal Display Television Format". In this chapter the authors explore a new display format free from television history. From their various experiments the authors form an overall preference for a 75 Hz interlaced scanning format.The appendices deal with the TGST de-interlacer, and is largely drawn from the authors' paper "Advanced Motion Estimation and Motion Compensated Deinterlacing" in which they combine two well-known algorithms.I found the book very enlightening. The authors do a good job of tying the various papers from which this book was formed into a cohesive format. Although there is a heavy and necessary use of mathematics, their writing style is very accessible and clear. If you are not sure you want to spend so much money on one book, you might try downloading off the web the papers I mentioned in this review. It will give you a good idea of the book's content and let you know if you like the authors' writing style. Personally, I highly recommend this book.
