Validation of Image Defect Models for Optical Character Recognition

Authors:
Yanhong Li;Daniel Lopresti;George Nagy;Andrew Tomkins
Affiliations:
-;-;-;-
Venue:
IEEE Transactions on Pattern Analysis and Machine Intelligence
Year:
1996

Citing 3
Cited 8

Ranking algorithms

Information retrieval
The String-to-String Correction Problem

Journal of the ACM (JACM)
Spatial sampling effects in optical character recognition

ICDAR '95 Proceedings of the Third International Conference on Document Analysis and Recognition (Volume 1) - Volume 1

Spatial Sampling of Printed Patterns

IEEE Transactions on Pattern Analysis and Machine Intelligence
Twenty Years of Document Image Analysis in PAMI

IEEE Transactions on Pattern Analysis and Machine Intelligence
A Statistical, Nonparametric Methodology for Document Degradation Model Validation

IEEE Transactions on Pattern Analysis and Machine Intelligence
Estimating Degradation Model Parameters Using Neighborhood Pattern Distributions: An Optimization Approach

IEEE Transactions on Pattern Analysis and Machine Intelligence
Effect of OCR error correction on Arabic retrieval

Information Retrieval
Language independent thresholding optimization using a Gaussian mixture modelling of the character shapes

Proceedings of the International Workshop on Multilingual OCR
Generation of training database using a noise model for OCR systems
An efficient parametrization of character degradation model for semi-synthetic image generation

Proceedings of the 2nd International Workshop on Historical Document Imaging and Processing

Quantified Score

Hi-index	0.14

Visualization

Abstract

In this paper, we consider the problem of evaluating character image generators that model distortions encountered in optical character recognition (OCR). While a number of such defect models have been proposed, the contention that they produce the desired result is typically argued in an ad hoc and informal way. We introduce a rigorous and more pragmatic definition of when a model is accurate: we say a defect model is validated if the OCR errors induced by the model are indistinguishable from the errors encountered when using real scanned documents. We describe four measures to quantify this similarity, and compare and contrast them using over ten million scanned and synthesized characters in three fonts. The measures differentiate effectively between different fonts and different scans of the same font regardless of the underlying text.