Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

Selecting a restoration technique to minimize OCR error.

M Cannon¹, M Fugate, D R Hush

¹Comput. Res. and Applications Group, Los Alamos Nat. Lab., NM, USA.

IEEE Transactions on Neural Networks

|February 2, 2008

Summary

This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Climate change and health: the role of health professionals.

Irish journal of psychological medicine·2024

Same author

Representations of hallucinations and dissociation in young adult literature: using literature to challenge stigma about psychosis.

Irish journal of psychological medicine·2022

Same author

The importance of anthropology in psychiatry.

Irish journal of psychological medicine·2022

Same author

Childhood adversity and recurrence of psychotic experiences during adolescence: the role of mediation in an analysis of a population-based longitudinal cohort study.

Psychological medicine·2022

Same author

Psychotic-like experiences? Trajectories and typologies of hallucinations and delusions from early adolescence to early adulthood in a population-based sample of Irish youth.

Irish journal of psychological medicine·2021

Same author

Student Mental Health and Wellbeing: Overview and Future Directions - Erratum.

Irish journal of psychological medicine·2021

Same journal

Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations.

IEEE transactions on neural networks·2013

Same journal

Guest editorial: special section on white box nonlinear prediction models.

IEEE transactions on neural networks·2011

Same journal

Data-based fault-tolerant control of high-speed trains with traction/braking notch nonlinearities and actuator failures.

IEEE transactions on neural networks·2011

Same journal

Guest editorial: special section on data-based control, modeling, and optimization.

IEEE transactions on neural networks·2011

Same journal

Neural network-based multiple robot simultaneous localization and mapping.

IEEE transactions on neural networks·2011

Same journal

Data-driven model-free adaptive control for a class of MIMO nonlinear discrete-time systems.

IEEE transactions on neural networks·2011

See all related articles

This study introduces a new learning method to improve optical character recognition (OCR) accuracy in document conversion. The generalized multiclass ratchet (GMR) algorithm effectively minimizes OCR errors for printed documents.

Area of Science:

Computer Science
Machine Learning
Document Analysis

Background:

Optical Character Recognition (OCR) is crucial for digitizing printed documents.
Minimizing OCR errors in document conversion remains a significant challenge.
Existing methods often struggle with complex document layouts and varied print qualities.

Purpose of the Study:

To develop a novel learning framework for document restoration.
To minimize average optical character recognition error in converting printed documents to ASCII text.
To introduce and analyze the generalized multiclass ratchet (GMR) algorithm.

Main Methods:

Derivation of an optimal function for mapping documents to restoration techniques.
Development of a nonparametric nearest-neighbor method.

Related Experiment Videos

Implementation of a direct solution via empirical error minimization.

Introduction of the generalized multiclass ratchet (GMR) algorithm, extending Kesler's construction and modifying the Pocket algorithm.

Main Results:

A finite sample bound on estimation error, independent of distribution, was proven for the direct method.
The empirical error minimization problem was shown to be an extension of traditional M-class classification.
The GMR algorithm was proven to asymptotically produce an optimal function.
Experimental results on document collections demonstrated the effectiveness of the proposed methods.

Conclusions:

The proposed learning framework and GMR algorithm offer a robust solution for reducing OCR errors.
The methods are computationally analyzed, with theoretical guarantees on performance.
This work advances the field of document image analysis and machine learning for practical applications.