By Topic

A label error process for discrete relaxation

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Hancock, E.R. ; Rutherford Appleton Lab., SERC, Didcot, UK ; Kittler, J.

A novel concept is introduced which involves the use of a label error process in conjunction with iterative discrete relaxation procedures. This idea allows the discrete relaxation procedure to draw on knowledge of constraints in the form of a dictionary of consistent labelings. According to this representation, nonphysical labelings are illegal and are not admitted. It is the legalization of nonphysical labelings that degrades the capacity of the label process to model consistency. The deadlock problem which results from the need to estimate the probability of nonphysical labelings is overcome by drawing on the idea of label corruption. Under certain nonrestrictive assumptions. the label corrupting process can be modeled by a binomial distribution of label errors. The number of such errors can be measured by the congruency between dictionary items and inconsistent labelings. Under the assumption of small label error probability, a model of the label error process has been demonstrated that can be realized efficiently by table lookup. The realization of the new methodology has been demonstrated for an edge-labeling application

Published in:

Pattern Recognition, 1990. Proceedings., 10th International Conference on  (Volume:i )

Date of Conference:

16-21 Jun 1990