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

Effects of opioid-free anesthesia on postoperative anxiety in patients undergoing modified radical mastectomy for breast cancer: a randomized controlled trial.

BMC anesthesiology·2026

Same author

SIRT4 Alleviates Retinal Ischemia-Reperfusion Injury Via Mediating Astrocytes Lipid Metabolism and Mitochondrial Function.

Investigative ophthalmology & visual science·2026

Same author

Automated camouflage pattern design based on conditional generative adversarial network and image quilting.

Scientific reports·2026

Same author

Volatile Oil from Magnolia sieboldii Alleviates Depres-sion and Regulates Neurotransmitter Expression in Reserpine-Induced Depressed Mice.

Journal of oleo science·2026

Same author

Application of Label-Free Detection Using Tapered Optical Fiber System for Head and Neck Cancer and Infectious Biomolecules.

Head & neck·2026

Same author

Angularly robust terahertz third-harmonic generation empowered by all-dielectric nonlocal metasurfaces.

Nanoscale·2026

Same journal

RETRACTED: Al-Hussain et al. Application of New Sodium Vinyl Sulfonate-co-2-Acrylamido-2-me[thylpropane Sulfonic Acid Sodium Salt-Magnetite Cryogel Nanocomposites for Fast Methylene Blue Removal from Industrial Waste Water. <i>Nanomaterials</i> 2018, <i>8</i>, 878.

Nanomaterials (Basel, Switzerland)·2026

Same journal

Correction: Jiang et al. Methods for Obtaining One Single Larmor Frequency, Either <i>v</i><sub>1</sub> or <i>v</i><sub>2</sub>, in the Coherent Spin Dynamics of Colloidal Quantum Dots. <i>Nanomaterials</i> 2023, <i>13</i>, 2006.

Nanomaterials (Basel, Switzerland)·2026

Same journal

Correction: Ekman et al. Synthesis, Characterization, and Adsorption Properties of Nitrogen-Doped Nanoporous Biochar: Efficient Removal of Reactive Orange 16 Dye and Colorful Effluents. <i>Nanomaterials</i> 2023, <i>13</i>, 2045.

Nanomaterials (Basel, Switzerland)·2026

Same journal

Ti<sub>3</sub>C<sub>2</sub>T<sub>x</sub>-Based Materials and Coatings for De-Icing and Defogging of Wind Turbine Blades: Materials Basis, Structural Design, Engineering Integration, and Future Opportunities.

Nanomaterials (Basel, Switzerland)·2026

Same journal

Influence of the Ripeness Stages of the Precursors on the Optical Characteristics of Carbon Dots Obtained from Valencia Orange Peels (<i>Citrus sinensis</i> L. Osbeck) by Hydrothermal Synthesis.

Nanomaterials (Basel, Switzerland)·2026

Same journal

Insights into ALD Growth of Al-Based Dielectric Stack on 4H-SiC.

Nanomaterials (Basel, Switzerland)·2026

See all related articles

Search research articles

Related Experiment Video

Updated: Jun 6, 2025

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

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

Published on: December 15, 2023

Metasurface-Based Image Classification Using Diffractive Deep Neural Network.

Kaiyang Cheng¹, Cong Deng¹, Fengyu Ye¹

¹International School of Microelectronics, Dongguan University of Technology, Dongguan 523808, China.

Nanomaterials (Basel, Switzerland)

|November 26, 2024

Summary

This summary is machine-generated.

This study introduces a diffractive deep neural network (D2NN) using all-dielectric metasurfaces for photonic computing. This AI-driven approach achieves over 90% accuracy in handwritten digit classification, enabling faster, more accurate optical computing.

Keywords:

diffractive deep neural network image classification metasurfaces

More Related Videos

Deep Neural Networks for Image-Based Dietary Assessment

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

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

Related Experiment Videos

Last Updated: Jun 6, 2025

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

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

Published on: December 15, 2023

Deep Neural Networks for Image-Based Dietary Assessment

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

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

Area of Science:

Photonics and Artificial Intelligence
Metasurface-based optical computing

Background:

Traditional photonic computing faces limitations in fabrication and flexibility for data-driven applications.
Artificial intelligence algorithms accelerate photonic computing development but require adaptable modulation methods.

Purpose of the Study:

To propose a novel diffractive deep neural network (D2NN) framework for optical computing.
To demonstrate a flexible, all-dielectric metasurface for light modulation.
To achieve high accuracy in image classification tasks using photonic neural networks.

Main Methods:

Developed a D2NN framework utilizing a three-layer all-dielectric phased transmitarray.
Engineered silicon nanodisk meta-atoms to control phase profiles and maintain high transmittance (0.9 at 600 nm).
Mimicked a fully connected neural network using phase-only metasurfaces with 1024 units per layer.

Main Results:

Achieved over 90% accuracy in classifying handwritten digits ('0' to '5') using the D2NN framework.
Validated performance through full-wave simulations.
Demonstrated successful classification of more complex animal images by increasing network connectivity.

Conclusions:

The proposed D2NN framework offers a viable solution for flexible light modulation in photonic computing.
This all-optical computing approach shows potential for practical applications in image processing and machine vision.
Metasurface-based neural networks provide a path towards efficient and compact optical computation.