Institute News

<Go to News list

IEEE Distinguished lectures in Tampere and Helsinki

IEEE Signal Processing / Circuits and Systems Chapter of Finland, Tampere International Center for Signal Processing and Signal Processing Laboratory at HUT invite you to attend two IEEE Distinguished Lectures given by Professor Jar-Ferr Yang from National Cheng Kung University, Taiwan and Distinguished Lecturer of the IEEE CAS Society.

1. Fast Realization of Video Coders based on Characteristics of DCT Coefficients

Time: Wednesday, 15th of September, 13:00h. Place: Tampere University of Technology (TUT), Department of Information Technology, Korkeakoulunkatu 1, Tietotalo building, auditorium TB224, Tampere (Hervanta).

2. Compact and Efficient Recursive Structures for Realization of Discrete Sinusoidal Transforms

Time: Friday, 17th of September, 10:00h. Place: Helsinki University of Technology (HUT), Department of Electrical and Communications Engineering, Signal Processing Laboratory, Otakaari 5 A, Espoo, Lecture Room H302.

The lectures are free and open for everybody.

Welcome!

-------------

Lecture Abstracts

Lecture #1: Fast Realization of Video Coders based on Characteristics of DCT Coefficients

In video coders, the discrete cosine transform (DCT) is a kernel function for compression of spatial data redundancy. It is well-known that the DCT coefficients achieve the best representation of image in terms of energy compaction and computation complexity. In the video decoder, we can easily receive the quantized DCT coefficients, which are mostly zeros, after the lossless decoder. In this lecture, the materials will cover our 6 journal papers. In realizing inverse DCT (IDCT), we dont need to perform whole fast IDCT computation for those zero DCT coefficients. The fast implementation of the IDCT processing will be first discussed by using the coefficient-by-coefficient schemes. Besides, the quantized DCT coefficients provide valuable information for video post-processing, which can remove the blocky effect resulting from heavy data quantization. Then, the DCT-based adaptive postprocessors will be addressed. As to the encoder, we also can predict zero coefficients in the DCT transformation especially for the residual after motion estimation. The computation for motion estimation, DCT, and IDCT in the encoding loop could be reduced once the prediction algorithms of all-zero block after DCT quantization are developed.

Lecture #2: Compact and Efficient Recursive Structures for Realization of Discrete Sinusoidal Transforms

Discrete sinusoidal transforms, such as the discrete cosine transform (DCT), modified DCT (MDCT), discrete Fourier transform (DFT) are widely used in areas of signal analyses and signal compressions. Fast algorithms and efficient implementations of discrete sinusoidal transforms are the important tasks for real-time applications. Due great contribution of nanotechnologies, the speed of VLSI chips becomes faster and faster. For low cost consumer products, the compact and efficient realization of transforms in recursive structures has been activated recently. In this lecture, the design methodologies of recursive structures, which included our 6 published journal papers as well as recent conference papers, will be addressed. First, the compact and efficient recursive structure of the 1-D DCT, DFT, and MDCT will be first introduced. Then, the condensed recursive structures for 2-D DCT, 2-D IDCT, 2-D DST and 2-D IDST have been discussed. Finally, the recursive structure of multi-dimensional DCT, which could be used for video compression, will be presented. By proper pre-process and data permutation, all the discrete sinusoidal transforms can be decomposed into some condensed recursive kernels. Based on pre-addition of input data with the same transform base, the proposed recursive structures require fewer recursive loops than other algorithms. Due to compact recursive loops in selected recursive kernel, more accurate results and less power consumption can be easily achieved for mobile and portable information appliances.

--------------

About the lecturer

Jar-Ferr Yang was born in Keelung, Taiwan on September 15, 1954. He received the B. S. degree from the Chung-Yuan Christian University, Taiwan in 1977, and the M. S. degree from the National Taiwan University, Taiwan in 1979, and the Ph. D. degree from the University of Minnesota, Minneapolis, U. S. A. in 1988 all in electrical engineering. He was an instructor in the Chinese Naval Engineering School in 1979-1980 and he as an assistant researcher worked in the Data Transmission and Network Design Research Group, Telecommunication Laboratories, Taiwan during 1981-1984. From 1982 to 1984, he was an adjunct lecturer in the Chung-Yuan Christian University. From 1984 to 1988, he received the Government Study Abroad Scholarship supported his advanced study in the University of Minnesota. In 1988, he jointed the National Cheng Kung University started as an associate professor. He was the Chairman of the Center for Computer and Communication Research, National Cheng Kung University from 1997 to 2000. In 2002, he was a visiting scholar at the Department of Electrical Engineering, University of Washington. He has published over 60 journal and 90 conference papers and held 6 U.S. patents and 7 Taiwanese patents in areas of audio, speech, image, and video processing and compression. Currently, he is the Deputy Director of the Electrical and Information Technology Center and a professor at the Department of Electrical Engineering, National Cheng Kung University. His teaching and research primarily are in the areas of multimedia signal processing, multimedia compression, spectrum estimation, and neural networks