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 Videos

A deep learning framework for scientific chart data extraction and reconstruction.

Yang Yuan1,2, Sihan Liang1,2, Ji Zhang1,2

  • 1Computer Network Information Center, Chinese Academy of Sciences, Beijing, China.

Communications Engineering
|May 20, 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

MiR-335 inhibits migration of breast cancer cells through targeting oncoprotein c-Met.

Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine·2014
Same author

Dynamic ambulance reallocation for the reduction of ambulance response times using system status management.

The American journal of emergency medicine·2014
Same author

Sophorolipid production from biomass hydrolysates.

Applied biochemistry and biotechnology·2014
Same author

[Research on progress and prospect of kinase S6K1].

Sheng wu yi xue gong cheng xue za zhi = Journal of biomedical engineering = Shengwu yixue gongchengxue zazhi·2014
Same author

Amplified fluorescent aptasensor through catalytic recycling for highly sensitive detection of ochratoxin A.

Biosensors & bioelectronics·2014
Same author

A novel artificial immune algorithm for spatial clustering with obstacle constraint and its applications.

Computational intelligence and neuroscience·2014
Same journal

Mobile charges in MoS<sub>2</sub>/high-k oxide transistors: from abnormal instabilities to transient negative differential resistance.

Communications engineering·2026
Same journal

Bubble-raft inspired shape-assembly in flying robot swarm for uniform formation and obstacle traversal.

Communications engineering·2026
Same journal

Enhanced magnetic moment discrimination for multiplex nanoparticle quantification via dual-frequency nonlinearity probing.

Communications engineering·2026
Same journal

Ultrahigh-speed micromachining of sapphire by enhancing laser absorption.

Communications engineering·2026
Same journal

Industry-Academia Interface: Exploring the growth of Additive Manufacturing as an industry with Laura Del Río Fernández.

Communications engineering·2026
Same journal

Operating smart grids by customizing large model agents.

Communications engineering·2026
See all related articles

ChartRecover extracts quantitative data from scientific figures using deep learning, improving data accessibility and research transparency. This framework enables automated scientific verification and data mining for enhanced reproducibility.

Area of Science:

  • Computer Science
  • Data Science
  • Scientific Visualization

Background:

  • Scientific figures are crucial for empirical evidence and integrity assessment.
  • Extracting quantitative data from diverse scientific plots is challenging due to format heterogeneity and noise.
  • Advances in NLP have not fully addressed automatic quantitative data extraction from visual plots.

Purpose of the Study:

  • To present ChartRecover, a deep learning framework for end-to-end extraction and reconstruction of chart data from scientific figures.
  • To enable accurate quantitative data extraction from heterogeneous scientific figures.
  • To enhance machine understanding of scientific figures for downstream applications.

Main Methods:

  • Utilizing object detection to identify graphical elements across various plot types.

Related Experiment Videos

  • Implementing a coordinate transformation strategy for aligning pixel coordinates with numerical values via axis tick-mark alignment and adaptive conversion.
  • Benchmarking ChartRecover across diverse figure styles and perturbation conditions.

    • Main Results:

    • ChartRecover demonstrates strong generalization capabilities.
    • The framework achieves high-fidelity recovery of quantitative data from scientific figures.
    • Successful extraction of structured data for applications like structure-property relationship mining and automated scientific verification.

    Conclusions:

    • ChartRecover advances machine understanding of scientific figures.
    • The framework provides a scalable tool to enhance research transparency, reproducibility, and data accessibility.
    • Automated quantitative data extraction from scientific figures is feasible and beneficial for scientific discovery.