The University of Edinburgh -
Division of Informatics
Forrest Hill & 80 South Bridge

Research Paper #542

Title:Computing with Uncertainty: Intervals Versus Probabilities
Authors:Orr,M; Fisher,RB; Hallam,JC
Date: 1991
Presented:Shorter version accepted for publication in Proceedins of British Machine Vision Association Conference (BMVC'91).
Abstract:We compare two well known methods of computing with uncertain quantities as used for geometric reasoning in robotics and computer vision. One method represents errors with intervals and manipulates them using techniques from interval arithmetic and network relaxation. The other method uses normal probability distributions for representation and manipulates them with techniques from statistical estimation theory. We find that the method based on probabilities is better in terms of both speed and accuracy.

[Search These Pages] [DAI Home Page] [Comment]