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 Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Knowledge Distillation in Object Detection: A Survey from CNN to Transformer.

Sensors (Basel, Switzerland)·2026
Same author

Semi-Supervised Object Detection: A Survey on Progress from CNN to Transformer.

Sensors (Basel, Switzerland)·2026
Same author

Object Detection with Transformers: A Review.

Sensors (Basel, Switzerland)·2025
Same author

Scene flow based deep network for hand reconstruction using depth images.

Scientific reports·2025
Same author

EventEgo3D++: 3D Human Motion Capture from A Head-Mounted Event Camera.

International journal of computer vision·2025
Same author

Enhancing robustness and generalization in microbiological few-shot detection through synthetic data generation and contrastive learning.

Computers in biology and medicine·2025

Related Experiment Video

Updated: Oct 16, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

690

CasTabDetectoRS: Cascade Network for Table Detection in Document Images with Recursive Feature Pyramid and Switchable

Khurram Azeem Hashmi1,2,3, Alain Pagani3, Marcus Liwicki4

  • 1Department of Computer Science, Technical University of Kaiserslautern, 67663 Kaiserslautern, Germany.

Journal of Imaging
|October 22, 2021
PubMed
Summary

We developed CasTabDetectoRS, a novel framework for detecting tables in document images. This efficient method achieves state-of-the-art results on multiple datasets without complex pre-processing or heavy networks.

Keywords:
atrous convolutioncascade Mask R-CNNcomputer visiondeep neural networksdocument image analysisobject detectionrecursive feature pyramid networkstable detectiontable recognition

More Related Videos

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

2.0K
Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

9.5K

Related Experiment Videos

Last Updated: Oct 16, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

690
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

2.0K
Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

9.5K

Area of Science:

  • Computer Vision
  • Document Image Analysis

Background:

  • Table detection is crucial for information extraction from scanned documents.
  • Existing methods often require complex pre-processing, heavy networks, or memory-intensive operations.

Purpose of the Study:

  • To introduce CasTabDetectoRS, an efficient and effective end-to-end table detection framework.
  • To demonstrate superior performance without relying on computationally expensive components.

Main Methods:

  • CasTabDetectoRS utilizes Cascade Mask R-CNN with Recursive Feature Pyramid and Switchable Atrous Convolution.
  • A lightweight ResNet-50 backbone is employed, avoiding heavier networks and deformable convolutions.
  • The framework is trained end-to-end without pre- or post-processing steps.

Main Results:

  • CasTabDetectoRS achieved state-of-the-art results on four datasets: ICDAR-19, TableBank, UNLV, and Marmot.
  • Significant relative error reductions were observed: 56.36% (ICDAR-19), 20% (TableBank), 4.5% (UNLV), and 3.5% (Marmot).
  • Comparable results were obtained on the ICDAR-17 POD dataset.

Conclusions:

  • CasTabDetectoRS offers a highly effective and efficient solution for table detection in document images.
  • The framework demonstrates strong generalization capabilities through cross-dataset evaluations.
  • This work sets a new benchmark for table detection performance and efficiency.