BBSRC Portfolio Analyser
Award details
Inference and learning in machine vision
Reference
BBS/E/S/00000451
Principal Investigator / Supervisor
Professor J Marchant
Co-Investigators /
Co-Supervisors
Institution
Silsoe Research Institute
Department
Silsoe Research Institute Department
Funding type
Research
Value (£)
481,300
Status
Completed
Type
Institute Project
Start date
01/04/2001
End date
31/03/2005
Duration
48 months
Abstract
The objective is to investigate soft methods in machine vision which will include (a) developing methods that operate by training (b) developing algorithms for extracting approximate but useful information (c) developing methods for combining information from algorithms. Most methods in machine vision use precise mathematical techniques to make some sort of measurement. As machine vision matures, and therefore is expected to deliver real working systems, we see a growing strand of research towards the softer end of computing ie systems that can learn and systems that can reason and come to conclusions given various sources of information. This movement may well be encouraged by the fact that humans probably do not do hard calculations but nevertheless make reasonable decisions. The understanding generated by this project will be generic but in particular we will develop algorithms for colour, texture and 3D relationships. With natural objects it may also be easier to get qualitative rather than quantitative outputs and we will investigate to what degree the scene can be classified given this loose information. We will also investigate whether and by how much performance is improved by combining evidence from a series of loose algorithms.
Summary
unavailable
Committee
Closed Committee - Engineering & Biological Systems (EBS)
Research Topics
X – not assigned to a current Research Topic
Research Priority
X – Research Priority information not available
Research Initiative
X - not in an Initiative
Funding Scheme
X – not Funded via a specific Funding Scheme
I accept the
terms and conditions of use
(opens in new window)
export PDF file
back to list
new search