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

Gramicidin D as a Multi-Target Antimicrobial Against Staphylococcus aureus: Biofilm Disruption, Virulence Attenuation, Antibiotic Potentiation, and Protein-Peptide Docking Insight.

Probiotics and antimicrobial proteins·2026
Same author

Mechanistic evaluation of amorphous InGaZnO-based one transistor-one capacitor memory cell <i>via</i> fast current-voltage measurements and modeling.

Nanoscale·2026
Same author

Longitudinal Structural and Functional Trajectories Following Netarsudil Intensification or Surgical Escalation in Primary Open-Angle Glaucoma.

Seminars in ophthalmology·2026
Same author

A flap mechanics testbed for skin reconstructive surgery: Evaluating the mechanical interaction between flap design and anisotropy.

Acta biomaterialia·2026
Same author

Environmental drivers of influenza A and B: causal analysis and forecasting.

BMC public health·2026
Same author

Anomalous ultrafast lithium-ion transport through boron nitride nanotube membranes.

Nature nanotechnology·2026

Related Experiment Video

Updated: Jul 29, 2025

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping
07:11

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping

Published on: December 8, 2023

1.6K

Data Preprocessing and Augmentation Improved Visual Field Prediction of Recurrent Neural Network with Multi-Central

Jeong Rye Park1, Sangil Kim2, Taehyeong Kim2

  • 1Finance Fishery Manufacture Industrial Center on Big Data, Pusan National University, Busan, Republic of Korea.

Ophthalmic Research
|May 26, 2023
PubMed
Summary

Data preprocessing and augmentation significantly improved recurrent neural network (RNN) visual field (VF) prediction. Periodic RNN models demonstrated superior performance over aperiodic models for forecasting future VF changes.

Keywords:
AugmentationData preprocessingRecurrent neural networkVisual field prediction

More Related Videos

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

2.8K
Topographical Estimation of Visual Population Receptive Fields by fMRI
06:02

Topographical Estimation of Visual Population Receptive Fields by fMRI

Published on: February 3, 2015

9.3K

Related Experiment Videos

Last Updated: Jul 29, 2025

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping
07:11

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping

Published on: December 8, 2023

1.6K
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

2.8K
Topographical Estimation of Visual Population Receptive Fields by fMRI
06:02

Topographical Estimation of Visual Population Receptive Fields by fMRI

Published on: February 3, 2015

9.3K

Area of Science:

  • Ophthalmology
  • Medical Informatics
  • Artificial Intelligence

Background:

  • Glaucoma management relies on accurate visual field (VF) monitoring.
  • Recurrent neural networks (RNNs) show promise for predicting VF progression.
  • Multi-center datasets present challenges due to variable VF test intervals.

Purpose of the Study:

  • To evaluate if data preprocessing and augmentation enhance RNN-based VF prediction.
  • To compare periodic versus aperiodic RNN models for VF forecasting.
  • To assess the impact of different data augmentation strategies on prediction accuracy.

Main Methods:

  • Retrospective analysis of 331,691 VFs from five glaucoma services (2004-2021).
  • Application of data augmentation to create fixed-interval datasets (365 ± 60 days and 180 ± 60 days).
  • Utilized RNNs with varying long- and short-term memory (LSTM) cells, comparing 5 vs. 6 LSTMs and periodic vs. aperiodic models.

Main Results:

  • Periodic RNN models (D=365) significantly outperformed aperiodic models (MAE: 2.56 vs. 3.26 dB, p < 0.001).
  • Higher perimetric frequency (shorter intervals) improved prediction accuracy (RMSE: 3.15 vs. 3.42 dB for D=180 vs. D=365).
  • Increasing input VFs and using 6-LSTM cells in the D=180 periodic model enhanced robustness and prediction accuracy.

Conclusions:

  • Data preprocessing with augmentation is effective in improving RNN-based VF prediction.
  • Periodic RNN models offer significantly better future VF prediction compared to aperiodic models.
  • Optimized data augmentation and model architecture enhance the reliability of AI in glaucoma monitoring.