Vector Signal Processors in Data Compression and Image Processing

King, G (1990). Vector Signal Processors in Data Compression and Image Processing. MPhil thesis The Open University.

DOI: https://doi.org/10.21954/ou.ro.0000fc5f

Abstract

The objective is to evaluate the applicability of the Vector Signal Processor to real time signal processing for data compression or manipulation. Particular emphasis has been placed on its role as a co-processor and the contribution that it might be expected to make during joint activities with the host.

These activities would have the combination used as the embedded computing subsystem of a FAX machine or as an image processing unit in desk top publishing. In these cases the hypothesis is that the Vector Signal Processor would act as an accelerator for many computationally intensive applicable processes.

After a review of current data compression techniques and of specialised architectures which may also be appropriate it is concluded that the Vector Signal Processor is the best option available. The operational details are then discussed. In order to be able to approximately compare experimental results with other workers a benchmarking exercise is undertaken.

Following this is the core of the study which details schemes for data compression of data sources involving character symbols, line drawings, and grey scale pictures. This involves pattern matching and substitution,Transform coding and quadtrees.

New encoding procedures are suggested based on Morse code for the secondary encoding of symbols and on Delta modulation for quadtrees. Image entity manipulation is discussed followed by some speculative work on neural networks and error control coding.

It is concluded that some processes are well served by the Vector Signal Processor but that the lack of conditional decision making and the difficulty of performing certain arithmetic functions make the processor unwieldy in its necessary host interactions.

Viewing alternatives

Download history

Metrics

Public Attention

Altmetrics from Altmetric

Number of Citations

Citations from Dimensions

Item Actions

Export

About

Recommendations