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

Vision01:24

Vision

52.2K
Vision is the result of light being detected and transduced into neural signals by the retina of the eye. This information is then further analyzed and interpreted by the brain. First, light enters the front of the eye and is focused by the cornea and lens onto the retina—a thin sheet of neural tissue lining the back of the eye. Because of refraction through the convex lens of the eye, images are projected onto the retina upside-down and reversed.
52.2K

You might also read

Related Articles

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

Sort by
Same author

Stacked graph attention network hypertuned by Heteroscedastic and Evolutionary Bayesian Optimization for fault identification on nuclear power plants robust to sensor drift.

Scientific reports·2026
Same author

Diffusion-augmented YOLO26-Swin cascaded framework with hybrid SHAP-CAM for autonomous power grid inspection.

Autonomous intelligent systems·2026
Same author

Audit-as-code: a policy-as-code framework for continuous AI assurance.

Frontiers in artificial intelligence·2026
Same author

Audio-Based Engine Fault Diagnosis with Wavelet, Markov Blanket, ROCKET, and Optimized Machine Learning Classifiers.

Sensors (Basel, Switzerland)·2024
Same author

Interference recommendation for the pump sizing process in progressive cavity pumps using graph neural networks.

Scientific reports·2023
Same author

Video-Based Human Activity Recognition Using Deep Learning Approaches.

Sensors (Basel, Switzerland)·2023

Related Experiment Video

Updated: May 10, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

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

Published on: December 15, 2023

430

The shallowest transparent and interpretable deep neural network for image recognition.

Gurmail Singh1, Stefano Frizzo Stefenon2, Kin-Choong Yow3

  • 1Department of Computer Sciences, University of Wisconsin-Madison, Madison, WI, 53706, USA. gurmail.singh@wisc.edu.

Scientific Reports
|April 22, 2025
PubMed
Summary

A new transparent deep learning model, Shallow-ProtoPNet, offers interpretable decisions without black-box components. Its shallow, two-layer design ensures smaller size and suitability for embedded systems.

Keywords:
Deep learningImage classificationInterpretable modelsPrototypical part network

More Related Videos

Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

8.9K
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

1.7K

Related Experiment Videos

Last Updated: May 10, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

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

Published on: December 15, 2023

430
Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

8.9K
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

1.7K

Area of Science:

  • Artificial Intelligence
  • Computer Vision
  • Medical Imaging

Background:

  • Deep learning models require transparency for high-risk decision-making.
  • Existing interpretable models often rely on black-box components, limiting full transparency.

Purpose of the Study:

  • Introduce Shallow-ProtoPNet, a fully transparent deep learning model.
  • Address the need for explainable AI in critical applications.

Main Methods:

  • Developed Shallow-ProtoPNet with a transparent prototype layer and a fully connected layer.
  • Avoided using black-box convolutional layers as a baseline, differentiating from ProtoPNet.
  • Evaluated performance on X-ray image datasets.

Main Results:

  • Shallow-ProtoPNet demonstrated comparable performance to less transparent interpretable models.
  • The model's shallow architecture (one convolutional, one fully connected layer) results in a significantly smaller size.
  • Achieved full transparency by excluding black-box components.

Conclusions:

  • Shallow-ProtoPNet offers a fully transparent and efficient deep learning solution.
  • Its reduced size makes it ideal for deployment in embedded systems.
  • Provides a viable alternative for interpretable AI in sensitive domains.