Plenary Lecture

Development of Embedded in GPU Parallel Structures of Algorithms for Real Time Video and Audio Information Processing

Professor Alexander Bekiarski
Faculty of Telecommunications
Technical University Sofia

Abstract: There are many existing methods and solutions for parallel data processing like digital signal processors, programmable logic, specialized or custom integrated circuits, etc., but the creation of graphical processing unit (GPU) make them attractive for applications requiring real time data processing. Graphical processing units (GPU) are devices designed to exploit parallel shared memory based on floating point computations. They provide memory access speeds superior to those in conventional personal computer CPU-based systems. The features to update in parallel the data (usually video and audio information) and the execution of operations (usually time consuming mathematical operations) are first offered and applied in computer vision and games applications. The advances of graphical processing unit (GPU) for powerful parallel processing appear first testing them for computer vision and computer games applications. This lead to development of a programming platform CUDA (Compute Unified Device Architecture) to help and facilitate the developments in computer vision and computer games algorithms and embedding them in graphical processing unit (GPU) working with the high level parallel programming languages like C, C++, C##, Java, OpenGL and other high level programming languages. Here is proposed to exploit the ability of parallel processing and the high-speed memory access of graphical processing units (GPU) for development of embedded in GPU parallel structures of algorithms for real time video and audio information processing. The proposed parallel structures are based on existing and well working algorithms for video and audio information processing, which are modified to suit the specific internal graphical processing unit (GPU) structure and to embedded them in GPU parallel structures using the a programming platform CUDA (Compute Unified Device Architecture). Many test are carried out with the developed parallel structures as corresponding video and audio information processing examples and their work are presented here as suitable experimental results in comparison their classical work without using GPU parallel structures and programming platform CUDA.

Brief Biography of the Speaker: Born in 1944, Plovdiv, Bulgaria. He received M.S. degree in Communications in 1969 in Technical University, Sofia. Ph. D in Television and Image Processing in 1975, Assoc. Prof. since 1987 in the same University. Proffesor since 2010 in Technical University-Sofia University.Vice-Dean of Faculty on Life-Long Learning Center since 2005, Vice-Dean of French Language Faculty of Electrical Engineering since 2006. The author over 216 research papers in Image Processing Systems, Pattern Recognitions, Neural Networks etc. Currently the leader of courses in Basic of Television, Television Systems, Theory of Coding, Digital Signal Processors etc. His scientific iterests encompass Video and Audio Processing, Digital TV, Neural Networks, Artificial Intelligence in Video and Audio, Artificial Intelligence Programming Languages Lisp Prolog, Expert Systems, Robotics Camera Eye and Microphone Arrays, Signal Processors, Embedded Systems, Microcontrollers, Programming Languages C++, Java, Matlab etc.

Bulletin Board


The Conference Program is online.

The Conference Guide is online.

The paper submission deadline has expired. Please choose a future conference to submit your paper.

Plenary Speakers

WSEAS Main Site

Publication Ethics and Malpractice Statement