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: Nov 11, 2025

Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging
11:38

Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging

Published on: October 4, 2024

843

Generative adversarial network based data augmentation to improve cervical cell classification model.

Suxiang Yu1, Shuai Zhang2, Bin Wang1

  • 1Department of Pathology, The Fourth Central Hospital of Baoding City, Baoding 072350, China.

Mathematical Biosciences and Engineering : MBE
|March 24, 2021
PubMed
Summary

Related Concept Videos

Methods of Nuclear Reprogramming01:24

Methods of Nuclear Reprogramming

2.0K
Nuclear reprogramming is a process of transforming one cell type into an unrelated cell type by epigenetic changes that alter the cell’s original gene expression pattern. Such epigenetic changes force cells to express a different set of genes, which play a significant role in inducing transformation into other cell types. Nuclear reprogramming offers applications in reproductive cloning for livestock propagation and regenerative medicine — developing patient-specific cells for...
2.0K

You might also read

Related Articles

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

Sort by
Same author

[Study of electroreflectance spectrum and Franz-Keldysh effect at metal-GaAs interfaces].

Guang pu xue yu guang pu fen xi = Guang pu·2008
Same author

[Study on electro-degradation of new conjugated polymer PFO-BT15 light emitting diodes].

Guang pu xue yu guang pu fen xi = Guang pu·2008
Same author

Comparison of the curative effects of video assisted thoracoscopic anterior correction and small incision, thoracotomic anterior correction for idiopathic thoracic scoliosis.

Chinese medical journal·2008
Same author

Distribution and sources of mercury in soils from former industrialized urban areas of Beijing, China.

Environmental monitoring and assessment·2008
Same author

[Main flavonoids from Sophora flavescenes].

Yao xue xue bao = Acta pharmaceutica Sinica·2008
Same author

External validation and prediction employing the predictive squared correlation coefficient test set activity mean vs training set activity mean.

Journal of chemical information and modeling·2008
Same journal

Modeling the impact of budget limitation on the screening and treatment pathway of HPV-induced precancerous cervical lesions.

Mathematical biosciences and engineering : MBE·2026
Same journal

Modeling the effects of trait-mediated dispersal on coexistence of two species: Competition and non-consumptive predator-prey.

Mathematical biosciences and engineering : MBE·2026
Same journal

A close look at the viral reduction rate in target cell limited models.

Mathematical biosciences and engineering : MBE·2026
Same journal

A stochastic agent-based model for simulating tumor-immune dynamics and evaluating therapeutic strategies.

Mathematical biosciences and engineering : MBE·2026
Same journal

Addressing domain shift via imbalance-aware domain adaptation in embryo development assessment.

Mathematical biosciences and engineering : MBE·2026
Same journal

Effect of drug resistance on an HIV epidemic in heterogeneous populations.

Mathematical biosciences and engineering : MBE·2026
See all related articles
This summary is machine-generated.

Early cervical cancer screening aids survival but burdens pathologists. A generative adversarial network (GAN) creates synthetic abnormal cells, improving convolutional neural network (CNN) classification accuracy for early detection.

Area of Science:

  • Computational pathology
  • Medical imaging analysis
  • Artificial intelligence in healthcare

Background:

  • Early cervical cancer screening significantly improves survival rates.
  • Pathologist workload for manual cervical cell screening is substantial.
  • Class imbalance and small sample sizes of abnormal cells pose challenges in automated classification.

Purpose of the Study:

  • To develop an automated cervical cell classification model to assist pathologists.
  • To address the challenges of small sample size and class imbalance in abnormal cervical cell detection.
  • To evaluate the effectiveness of generative adversarial networks (GANs) in augmenting training data for improved classification.

Main Methods:

  • A generative adversarial network (GAN) was trained on images of abnormal cervical cells to generate synthetic abnormal cell images.
Keywords:
cervical cell classificationdata augmentationdeep learninggenerative adversarial network (GAN)

Related Experiment Videos

Last Updated: Nov 11, 2025

Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging
11:38

Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging

Published on: October 4, 2024

843
  • A convolutional neural network (CNN) was trained using a combination of real and GAN-generated abnormal cell images.
  • Four experimental approaches were compared, including data under-sampling, transfer learning, direct training with generated data, and pre-training with generated data followed by fine-tuning.

    • Main Results:

    • GAN-generated abnormal cell images effectively mitigated the small sample and class imbalance problems in cervical cell classification.
    • The CNN model pre-trained with GAN-generated images and fine-tuned with real images achieved the highest performance.
    • The optimal model achieved an Area Under the Curve (AUC) value of 0.984.

    Conclusions:

    • Generative adversarial networks (GANs) provide a viable solution for data augmentation in imbalanced medical image datasets like cervical cytology.
    • Pre-training a CNN with synthetic data generated by GANs, followed by fine-tuning with real data, significantly enhances classification performance.
    • This approach holds promise for improving the accuracy and efficiency of automated cervical cancer screening.