Parallel architectures and parallel algorithms for integrated vision systems

Book · Mon Jan 01 00:00:00 EST 1990

OSTI ID:6468457

Choudhary, A N ^[1]; Patel, J H ^[2]

---

1. Syracuse Univ. (US)
2. Univ. of Illinois, IL (US)

Computer vision is one of the most complex and computationally intensive problems in computer science and electrical engineering. Hierarchical and partitionable architectures for vision systems have been suggested to provide the required flexibility, reconfigurability and partitionability to solve computer vision problems. This book examines important issues in designing these architectures to be adapted to perform a given computation efficiently while different computations are performed in other parts of the architecture. Important issues, such as how to map different algorithms and what techniques can be used to implement algorithms with different characteristics, are also addressed. The book develops a model of computation for problems in integrated vision systems (IVS). The model then examines and incorporates the computation, communication and other requirements for parallel implementation of integrated vision systems. Using these requirements, a criterion is developed to design hierarchical and partitionable multiprocessor architectures for IVS's. This book develops several techniques for mapping, scheduling and load balancing of IVS algorithms. These techniques are general enough to be applied to many IVS's and can be used on any MIMD shared or distributed memory machines.

• OSTI ID: 6468457
• Country of Publication: United States
• Language: English

Related Subjects

99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE
COMPUTER GRAPHICS
PARALLEL PROCESSING
IMAGE PROCESSING
ALGORITHMS
ARRAY PROCESSORS
COMPUTER ARCHITECTURE
SUPERCOMPUTERS
COMPUTERS
DIGITAL COMPUTERS
MATHEMATICAL LOGIC
PROCESSING
PROGRAMMING
990200* - Mathematics & Computers