Skip to main navigation Skip to search Skip to main content

Reading Text on Both Sides of a Sheet of Paper with a Neural Network Classifier

  • Haolian Shi
  • , Rahul Komatinenir
  • , Yuan Max
  • , Cédric Pradalier
  • , Leor Jacobi
  • , Yikun Jiang
  • , Alexandre Locquet
  • , D. S. Citrin
  • GT-CNRS UMI 2958
  • Georgia Institute of Technology
  • Peking University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This work explores the imaging of characters written in iron gall ink on both sides of a sheet of paper using THz-TDS measurements. A convolutional neural network is trained to classify each pixel, providing an estimation of text presence on both sides. Experimental validation is conducted on single- and two-sheet samples, with limitations observed in the readability of the back side.

Original languageEnglish
Title of host publication50th International Conference on Infrared, Millimeter, and Terahertz Waves, IRMMW-THz 2025
PublisherIEEE Computer Society
ISBN (Electronic)9798350378832
DOIs
StatePublished - 2025
Event50th International Conference on Infrared, Millimeter, and Terahertz Waves, IRMMW-THz 2025 - Espoo, Finland
Duration: 17 Aug 202522 Aug 2025

Publication series

NameInternational Conference on Infrared, Millimeter, and Terahertz Waves, IRMMW-THz
ISSN (Print)2162-2027
ISSN (Electronic)2162-2035

Conference

Conference50th International Conference on Infrared, Millimeter, and Terahertz Waves, IRMMW-THz 2025
Country/TerritoryFinland
CityEspoo
Period17/08/2522/08/25

Bibliographical note

Publisher Copyright:
© 2025 IEEE.

Fingerprint

Dive into the research topics of 'Reading Text on Both Sides of a Sheet of Paper with a Neural Network Classifier'. Together they form a unique fingerprint.

Cite this