Sciweavers

Share
CVPR
1998
IEEE

A Methodology for Deriving Probabilistic Correctness Measures from Recognizers

9 years 11 months ago
A Methodology for Deriving Probabilistic Correctness Measures from Recognizers
This paper describes the derivation of probability of correctness from scores assigned by most recognizers. Motivation for this research is three-fold: i probability values can be used to rerank the output of any recognizer by using a new set of training data; if the training data is su ciently large and representative of the test data, the recognition rates are seen to improve signi cantly, ii derivation of probability values puts the output of di erent recognizers on the same scale; this makes comparison across recognizers trivial, and iii word recognition can be readily extended to phrase and sentence recognition because the integration of language models becomes straightforward. We have conducted an extensive set of experiments. The results show a reranking of recognition choices based on the derived probability values leading to an enhancement in performance.
Djamel Bouchaffra, Venu Govindaraju, Sargur N. Sri
Added 12 Oct 2009
Updated 30 Oct 2009
Type Conference
Year 1998
Where CVPR
Authors Djamel Bouchaffra, Venu Govindaraju, Sargur N. Srihari
Comments (0)
books