Yakup Paker

A virtual studio for live broadcasting: the Mona Lisa project

The Virtual Studio concept replaces real background sets with computer-generated synthetic scenes. Image sequences combine real foreground action, shot in the studio, with computer generated backgrounds. The Mona Lisa project seeks to develop and integrate technologies (algorithms, software, and hardware) needed to construct, handle, and render 3D models in real time.

Optimal Scheduling Algorithms for Communication Constrained Parallel Processing

With the advent of digital TV and interactive multimedia over broadband networks, the need for high performance computing for broadcasting is stronger than ever. Processing a digital video sequence requires considerable computing. One of the ways to cope with the demands of video processing in real-time, we believe, is parallel processing. Scheduling plays an important role in parallel processing especially for video processing applications which are usually bounded by the data bandwidth of the transmission medium. Although periodic real-time scheduling algorithms have been under research for more than a decade, scheduling for continuous data streams and impact of scheduling on communication performance are still unexplored. In this paper we examine periodic real-time scheduling assuming that the application is communication constrained where input and output data sizes are not equal.

Customized television: standards compliant advanced digital television

This paper describes a European Union supported collaborative project called CustomTV based on the premise that future TV sets will provide all sorts of multimedia information and interactivity, as well as manage all such services according to each users or group of users preferences/profiles. We have demonstrated the potential of recent standards (MPEG-4 and MPEG-7) to implement such a scenario by building the following services: an advanced EPG, weather forecasting, and stock exchange/flight information.

A novel hexagonal search algorithm for fast block matching motion estimation

Based on real-world image sequence characteristics of center-biased motion vector distribution, a Hexagonal (HS) algorithm with center-biased checking point pattern for fast block motion estimation is proposed. The HS is compared with full search (FS), four-step search (4SS), new three-step search (NTSS), and recently proposed diamond search (DS) methods. Experimental results show that the proposed technique provides competitive performance with reduced computational complexity.

Adaptive parallelism under Equus

The authors describe adaptively parallel computations under Equus (T. Kindberg, 1990; A. V. Sahiner, 1991). These computations execute on a processor pool, and expand and contract as the number of processor nodes allocated to them varies over their run-time. They are based upon a hierarchical master-worker structure. The number of worker processes changes with the number of allocated nodes, and so does the number of processes that act as servers to them (such as the masters). The authors use an image-processing example to describe how workers and servers are added and withdrawn at run-time. The affected processes are synchronised, communication linkages are changed, and in some cases, state is transferred between them. Reconfigurations are transparent to worker (and other client) processes, but not all can be made transparent to servers.<<ETX>>

A study of concurrency in MPEG-4 video encoder

The traditional boundaries between the computer, TV/film industries and telecommunications, are blurring. Therein lies the focus of MPEG-4: the convergence of common applications of the above mentioned three industries. MPEG-4 aims to satisfy the new requirements and expectations, by providing an audio visual coding standard allowing for interactivity, high compression and universal accessibility. Moreover coding of video data is an important feature in digital industry and the MPEG-4 video encoder has to be able to encode efficiently visual information and probably in real time conditions. Therefore a lot of effort has been spent in real time video encoding and decoding issues using either hardware or software solutions. This paper addresses these issues by describing a parallel software implementation of the MPEG-4 video encoder using multithread techniques. A scheduling policy is proposed to guarantee via a buffer synchronisation a significant speed up which under some special circumstances reach an optimised load balancing solution. The proposed scheme is tested using the hardware resources of Unix multiprocessor hardware platform (running Solaris 2.0) by rewriting and modifying the current MPEG-4 encoder.

Matrix Factorizations for Parallel Integer Transformation

Integer mapping is critical for lossless source coding and has been used for multicomponent image compression in the new international image compression standard JPEG 2000. In this paper, starting from block factorizations for any nonsingular transform matrix, we introduce two types of parallel elementary reversible matrix (PERM) factorizations which are helpful for the parallelization of perfectly reversible integer transforms. With improved degree of parallelism and parallel performance, the cost of multiplications and additions can be, respectively, reduced to O(logN) and O(log2N) for an N by N transform matrix. These make PERM factorizations an effective means of developing parallel integer transforms for large matrices. We also present a scheme to block the matrix and allocate the load of processors for efficient transformation

A Parallel Accelerator for Using Synthetic Images in TV and Video Production

With the recent developments in digital media computers are used increasingly in TV production. Computer generated logos, animated objects and characters, and video wizardry generated with complex image processing operations are now part of programs we see daily on our television screens. A large proportion of the techniques employed for this purpose, for example, computer graphics techniques to achieve photo-realism, like ray tracing [1] and radiosity [2], and image transformation applications such as non-linear mapping [3] are computationally expensive and they require high performance computation engines. Multiprocessors provide a possible solution to this problem. Hardware platforms built using high-performance processors such as MIPS R4400, Intel i860, Transputers, and DSPs are already being used for this purpose [4,5].

Optimum Tactics of Parallel Multi-Grid Algorithm with Virtual Boundary Forecast Method Running on a Local Network with the PVM Platform

In this paper, an optimum tactic of multi-grid parallel algorithm with virtual boundary forecast method is disscussed, and a two-stage implementation is presented. The numerical results of solving a non-linear heat transfer equation show that the optimum implementation is much better than the non-optimum one.

A parallel FFT algorithm for transputer networks

This paper presents a parallel FFT algorithm suitable for MIMD machines. It has the properties of uniform granularity, the dominance of computation, and suitability to large data matrices. The algorithm has been implemented for a tree and cubic types of transputer network in Occam language. A set of performance figures has been derived for different size of networks and initial matrices.