Скачать или смотреть IEICE English Webinar : Data Compression: Principles and Applications - Tomohiko Uyematsu

IEICE English Webinar: Distinguished Lecturer Program Series – September 25, 2024

"Data Compression: Principles and Applications"
Lecturer: Prof. Tomohiko Uyematsu
Tokyo Institute of Technology

Biography:
Tomohiko Uyematsu received the B.E., M.E., and D.E. degrees from Tokyo Institute of Technology. He is currently professor of the School of Engineering, Tokyo Institute of Technology. In 2012 and 2020, He was appointed to Vice President and Dean, respectively in Tokyo Institute of Technology. He is now serving as President-Elect in IEICE. He was
Associate Editor of IEEE Trans. Information Theory.

Summary：
Data compression is a technique that is widely used in data storage and communication. In this presentation, we focus on the lossless data compression, which allows the original data to be completely restored, and explain the basic principles of data compression in information theory as well as practical coding algorithms. Specifically, we explain information sources such as memoryless sources and stationary ergodic sources, types of source codes, prefix codes, Kraft's inequality, and entropy. Then, we prove that the performance limit of data compression, determined by the average code length, is the entropy of the source. Further, we clarify how to construct prefix codes whose average code length differs from the entropy within 1 bit. We also explain practical coding algorithms, including Huffman codes that minimize the average code length, and Lempel-Ziv 77 and Lempel-Ziv 78 codes, which are universal codes whose average code length converges to the entropy as the length of input data becomes long, even if the statistical properties of the source are unknown. Finally, we investigate the actual performance of these codes by compressing test data.

Send Feedback:
https://forms.office.com/r/twx105xf5y

Follow us on Twitter: @Ieice_trans