What I’ve learned in this course
Guan-Ju Peng
Graduate Institute of Electronics Engineering, National Taiwan University
In this semester, I’ve read three books which are ”Video Codec Design”, ”H.264 and MPEG-4 Video Compression” and ”Video Processing and communications”. I also sur- vey several papers in which I think interesting are ”Bit allocation for dependent quanti- zation with applications” and ”Overview of the Scalable Video Coding Extension of the H.264/AVC Standard”. I will express what I’ve learned in the following.
1 Books
The first book I read is ”Video Codec Design”, which is a relatively easy reading book.
In this book, it first introduce the video format and the different compression standards.
It shows the tool set of each standard and discuss the advantage and disadvantage of the standards. The second part introduces the coding tools separately. They are motion estimation, transform coding, entropy coding, pre-, post- processing and rate-distortion problem. However, although the book describe how these tools work, it doesn’t show why they can work. For example, the discrete cosine transform and wavelet transform is introduced by a mathematical formula. But the reason why the signal after transformation can help the coding efficiency does not appear. I suggest a book ”Digital Coding of Waveform” which is an old book but has a good explanation from theory to practical.
The third part describes some implementation issues which may be useful while adopting the video codec into applications.
The second book I’ve read is ”H.264 and MPEG-4 Video Compression”. Since this book and the previous book are written by the same author, they are similar and repeat in many chapters. Thus only chapter 5 and 6 has more information different to the previous book. The book shows the tools and profiles of MPEG-4 and H.264/AVC.
However, the reader can also acquire the related information in the paper ”Overview of 1
the H.264/AVC Video Coding Standard”, in IEEE Transactions on Circuits and Systems for Video Technology” for details.
The third book is ”Video Processing and Communication”. Unlike the previous two books, this one shows many details on signal processing and is useful not only on video coding. In fact, the book is designed for Image/Video processing in origin. The books is full of materials, thus I only read half of the book. I planned to finish the reading in the future two months.
2 Papers
Since I am interested in the rate distortion problem in the video coding, I find a paper
”Bit allocation for dependent quantization with applications” which describe the problem elegantly. The main idea of the paper is to consider the distortion function universally.
Generally, when we consider the bit allocation problem, only current frame/block is con- sidered. However, in most of time, the rate-distortion is dependent on the previous frame/block because we usually utilize their dependency to increase the coding efficiency.
The paper proposes the rate-distortion curve considering all dependent frame/block to optimize the coding performance. But it also costs lots of time and power.
The other paper I’ve read is ”Overview of the Scalable Video Coding Extension of the H.264/AVC Standard”. The paper proposes several tools to enhance the coding efficiency while supporting scalability. Although the major problem which named as ”drafting effect” due to the prediction loop is not solved, the coding performance is approaching the single layer coding.
In this course, I’ve studied many concepts of Video/Image coding which help further research works. Prof. Tung gives a clear and efficient explanation on each topic even some of them are really complicated and hard to describe. I really appreciate everything I’ve learned in this course.
2
