Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Jun 16, 2026

Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

A large-scale benchmark dataset for Urdu script optical character recognition with systematic augmentation.

Fauzia Yasir1, Majida Kazmi1,2, Saad Ahmed Qazi1,2

  • 1Faculty of Electrical and Computer Engineering, NED University of Engineering and Technology, Karachi 75270, Pakistan.

Data in Brief
|June 15, 2026
PubMed
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

Binaural Beat Stimulation for Cognitive and Psychological Enhancement in Alzheimer's Disease: A Pilot Study.

IEEE transactions on neural systems and rehabilitation engineering : a publication of the IEEE Engineering in Medicine and Biology Society·2026
Same author

Binaural Beat Stimulation Enhances Cognitive Function in Alzheimer's Disease via Temporal Lobe Activation: An sLORETA Study.

Biomedicines·2026
Same author

Impact of binaural beat stimulation on working memory: a graph theory network approach.

Frontiers in neuroscience·2026
Same author

Modulated theta band frequency with binaural beat stimulation correlates with improved cognitive scores in Alzheimer's patients.

Frontiers in aging neuroscience·2025
Same author

Brain Insights and Resolution of Youth Depression through Neurotechnology.

Clinical EEG and neuroscience·2024
Same author

Predicting the effectiveness of binaural beats on working memory.

Neuroreport·2024
Same journal

Dataset of pesticide and trace metal concentrations in the pollen provisions of wild bees and surrounding soils across European bee hotels.

Data in brief·2026
Same journal

Dataset of structure-activity relationships in Pd/ZrO<sub>2</sub>-TiO<sub>2</sub> catalysts for furfural reductive amination: Batch vs Operando ATR-FTIR.

Data in brief·2026
Same journal

Dine in or take out dataset: user behavior in an interactive virtual reality café.

Data in brief·2026
Same journal

Heavy commercial vehicles' disposition: Anonymized dataset of German truck freight transport order trips (DT-DISPO).

Data in brief·2026
Same journal

A harmonized fast-fashion garment-variant dataset for textile circularity and sustainability assessment.

Data in brief·2026
Same journal

Terahertz reflectivity dataset: Reading text on both sides of the page.

Data in brief·2026
See all related articles

A new dataset, FIPU-OCR-CHAR, offers 337,680 printed Urdu characters for optical character recognition (OCR) research. This resource aims to improve Urdu script digitization and OCR model development for low-resource languages.

Area of Science:

  • Computer Vision and Pattern Recognition
  • Natural Language Processing
  • Digital Humanities

Background:

  • Urdu, spoken by over 230 million people, lacks sufficient digital resources, particularly for optical character recognition (OCR).
  • The variability of Urdu fonts presents a significant challenge for developing generalizable OCR systems.
  • A lack of standardized benchmarks impedes reproducible research in Urdu OCR.

Purpose of the Study:

  • To introduce FIPU-OCR-CHAR, a comprehensive benchmark dataset for printed Urdu character recognition.
  • To provide a standardized resource for developing and evaluating Urdu OCR technologies.
  • To facilitate research in font-invariant classification and low-resource script digitization.

Main Methods:

  • Systematic dataset construction pipeline involving font collection, character definition, image rendering, and augmentation.
Keywords:
Computer visionData annotationImage augmentationLanguage digitizationLow-resource languageMachine learningNatural language processingOCR text images

Related Experiment Videos

Last Updated: Jun 16, 2026

Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

  • Rendering of 48 Urdu character classes (38 alphabets, 10 numerals) from 201 distinct fonts.
  • Application of 34 augmentation techniques to generate 337,680 labeled 28x28 pixel PNG images, organized into training, validation, and testing splits.

    • Main Results:

    • Creation of FIPU-OCR-CHAR, a dataset comprising 337,680 labeled images of isolated printed Urdu characters.
    • Dataset includes 201 unique Urdu fonts, covering 48 character classes with extensive augmentation.
    • Provided baseline ResNet-34 model and clear documentation for immediate use and reproducibility.

    Conclusions:

    • FIPU-OCR-CHAR addresses the critical need for large-scale, diverse datasets in Urdu OCR.
    • The dataset is suitable for training robust OCR models, font-invariant classifiers, and transfer learning applications.
    • Enables benchmarking of advanced deep learning architectures like CNNs and Vision Transformers for low-resource scripts.