Implementation of VQ Algorithms on a Reconfigurable Array Processor
Abstract
Vector quantization is being widely used in image data compression applications due to the facts that it is capable of achieving fractional bit rates with reasonable complexity and that the decoding is a very simple table look-up scheme. In image encoding, a vector quantizer accepts a block of pixels and outputs an address of the best matching tile stored in a codebook. The matching algorithm requires a large number of basic arithmetic operations in typical applications. Since real time coding is required in many video applications, the need for dedicated processing architectures arises naturally. This paper investigates the mapping of VQ algorithms onto an array processor to achieve near real-time compression of video images.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1991
- Accession Number
- ADA236483
Entities
People
- K. S. Thyagarajan
- Thomas B. Henderson