Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

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

Full-tendon inferior nasal transposition of the lateral rectus muscle for partial third nerve palsy.

Taiwan journal of ophthalmology·2026

Same author

Associations of Dietary Patterns and Physical Activity with Sleep Quality and Metabolic Health Markers in Patients with Obstructive Sleep Apnea: An Exploratory Pilot Study.

Nutrients·2026

Same author

Corneal endothelial and ocular biometric parameters in children with history of retinopathy of prematurity and treated with laser or intravitreal injections.

Canadian journal of ophthalmology. Journal canadien d'ophtalmologie·2025

Same author

Frailty subtypes and white matter alterations in older adults without dementia.

Psychiatry and clinical neurosciences·2025

Same author

Dynamics of two distinct memory interactions during water seeking in <i>Drosophila</i>.

Proceedings of the National Academy of Sciences of the United States of America·2025

Same author

Measurements of Objective Cyclotorsion in a Population of Healthy Children.

Journal of ophthalmology·2024

Search research articles

Related Experiment Video

Updated: Oct 25, 2025

Assessing Pupil-linked Changes in Locus Coeruleus-mediated Arousal Elicited by Trigeminal Stimulation

Assessing Pupil-linked Changes in Locus Coeruleus-mediated Arousal Elicited by Trigeminal Stimulation

Published on: November 26, 2019

Pupil Size Prediction Techniques Based on Convolution Neural Network.

Allen Jong-Woei Whang¹, Yi-Yung Chen², Wei-Chieh Tseng¹

¹Department of Electronic and Computer Engineering, National Taiwan University of Science and Technology, Taipei City 106335, Taiwan.

Sensors (Basel, Switzerland)

|August 10, 2021

Summary

This study introduces a convolution neural network (CNN) algorithm for accurate pupil size measurement using ellipse parameters. The AI model achieves high performance on low-cost systems, enabling new applications in health and mental state monitoring.

Keywords:

biomedical imaging computational intelligence engineering in medicine and biology machine learning

More Related Videos

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

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

Published on: November 30, 2022

Pupillometry to Assess Auditory Sensation in Guinea Pigs

Pupillometry to Assess Auditory Sensation in Guinea Pigs

Published on: January 6, 2023

Related Experiment Videos

Last Updated: Oct 25, 2025

Assessing Pupil-linked Changes in Locus Coeruleus-mediated Arousal Elicited by Trigeminal Stimulation

Assessing Pupil-linked Changes in Locus Coeruleus-mediated Arousal Elicited by Trigeminal Stimulation

Published on: November 26, 2019

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

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

Published on: November 30, 2022

Pupillometry to Assess Auditory Sensation in Guinea Pigs

Pupillometry to Assess Auditory Sensation in Guinea Pigs

Published on: January 6, 2023

Area of Science:

Computer Vision
Artificial Intelligence
Biomedical Signal Processing

Background:

Pupil size is a key indicator of physiological and psychological states.
Existing AI research on pupils primarily focuses on eye-tracking, not size estimation.
Non-round pupil shapes necessitate advanced modeling beyond simple diameter measurements.

Purpose of the Study:

To develop a novel algorithm for calculating pupil size using a convolution neural network (CNN).
To represent pupil size using the major and minor axes of an ellipse, accommodating non-round shapes.
To optimize the CNN architecture for accuracy and efficiency in pupil size prediction.

Main Methods:

Utilized a CNN for pupil size calculation, with ellipse major and minor axes as network outputs.
Employed data augmentation and structural similarity calculations to mitigate overfitting in video-based datasets.
Investigated the impact of network depth and convolution filter field of view (FOV) on prediction accuracy.

Main Results:

Deepening the network and widening the FOV of convolution filters reduced mean error.
Achieved a mean error of 5.437% for pupil length and 10.57% for pupil area.
Demonstrated real-time performance (35 fps) on low-cost mobile embedded systems.

Conclusions:

The proposed CNN algorithm accurately estimates pupil size, even for non-round shapes.
Network architecture optimization (depth, FOV) is crucial for improving prediction accuracy.
The system's efficiency and low-cost potential open avenues for widespread application in health monitoring.