About Graham Tattersall

Home About Snape Signals Signal Processing Pattern Recognition Data Mining Research Papers

Overview:

Graham Tattersall’s career in the research, development, and application of data mining, signal and pattern processing spans thirty years. Early years were spent in the research and development departments of British Telecom and Nokia Electronics. Later he worked at the Universities of Keele and East Anglia. The last ten years have been spent as a consultant engineer and Honorary Senior Lecturer at the University of East Anglia.

Background:

During the 1970s Graham Tattersall worked on the development of digital telephone systems at British Telecom and Nokia. Much of this work focussed on the development of new types of digital filter structure for use with over-sampled ADCs. In the 1980’s he moved into academia; first at the University of Keele and then the University of East Anglia. As a Senior Lecturer at the University of East Anglia, he established a leading university research group specializing in artificial neural networks.

The group was particularly successful in developing sophisticated versions of the self-organising neural networks invented by Teuvo Kohonen. It was also a leader in developing n-tuple sampling networks as a powerful alternative to neural nets such as the MLP and RBF. Practical applications of the work on neural networks were aimed at speech processing and data mining, with particular emphasis on decision support systems.

More recently, he has concentrated on the development of data mining, signal, and pattern processing methods for commercial applications. Examples are the development and production of quality assured software tools for safety inspections of nuclear plant, and development of image processing tools for enhancement and interpretation of ultrasound images used in NDT of welds. An interesting aspect of the latter has been the development of beam-spread deconvolution algorithms that allow more precise defect location and sizing.

In the last three years he has worked as an industrial partner within a DTI funded project into personalization of HCI. His contribution to this project has centred on development of Bayesian and transform techniques for mining semantic content from document databases and analysing human user behaviour.

Home About Snape Signals Signal Processing Pattern Recognition Data Mining Research Papers