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

Quantifying Heat02:46

Quantifying Heat

57.1K
Thermal Energy Microscopically, thermal energy is the kinetic energy associated with the random motion of atoms and molecules. Temperature is a quantitative measure of “hot” or “cold”, which depends on the amount of thermal energy. When the atoms and molecules in an object are moving or vibrating quickly, they have a higher average kinetic energy (KE) (or higher thermal energy), and the object is perceived as “hot”, or it is described as being at a...
57.1K
Reducing Line Loss01:18

Reducing Line Loss

212
In a three-phase circuit, line loss is an indicator of energy dissipated as heat due to the resistance of transmission lines. To address this, incorporating transformers into the system—a step-up transformer at the source and a step-down transformer at the load—is a strategic solution. Two three-phase transformers are introduced to improve this.
With a step-up transformer at the source, the voltage is increased, thereby reducing the current in the transmission lines since power loss...
212
Deconvolution01:20

Deconvolution

271
Deconvolution, also known as inverse filtering, is the process of extracting the impulse response from known input and output signals. This technique is vital in scenarios where the system's characteristics are unknown, and they must be inferred from the observable signals.
Deconvolution involves several mathematical techniques to derive the impulse response. One common approach is polynomial division. In this method, the input and output sequences are treated as coefficients of...
271

You might also read

Related Articles

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

Sort by
Same author

Amodal Segmentation and Trait Extraction of On-Branch Soybean Pods with a Synthetic Dual-Mask Dataset.

Sensors (Basel, Switzerland)·2025
Same author

Driving-Related Cognitive Abilities Prediction Based on Transformer's Multimodal Fusion Framework.

Sensors (Basel, Switzerland)·2025
Same author

EasyDAM_V2: Efficient Data Labeling Method for Multishape, Cross-Species Fruit Detection.

Plant phenomics (Washington, D.C.)·2022
Same author

Deep-learning-based in-field citrus fruit detection and tracking.

Horticulture research·2022
Same author

Lightweight Fruit-Detection Algorithm for Edge Computing Applications.

Frontiers in plant science·2021
Same author

Easy domain adaptation method for filling the species gap in deep learning-based fruit detection.

Horticulture research·2021

Related Experiment Video

Updated: Sep 30, 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

666

A Pruning Method for Deep Convolutional Network Based on Heat Map Generation Metrics.

Wenli Zhang1, Ning Wang1, Kaizhen Chen1

  • 1Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China.

Sensors (Basel, Switzerland)
|March 10, 2022
PubMed
Summary
This summary is machine-generated.

This study introduces a novel deep convolutional network pruning method using heatmap-generated metrics to remove redundant features in infrared images. The alternating training and self-pruning strategy enhances model performance by reducing incorrect layer deletions.

Keywords:
layerspruning networkredundant features

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

9.4K

Related Experiment Videos

Last Updated: Sep 30, 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

666
Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

9.4K

Area of Science:

  • Computer Vision
  • Deep Learning
  • Machine Learning

Background:

  • Deep learning models excel at extracting features from visible images.
  • Infrared image analysis with deep networks is challenging due to homogeneous features and redundant information extraction.
  • Pruning network layers is crucial to mitigate redundant feature extraction in infrared image processing.

Purpose of the Study:

  • To propose an effective pruning method for deep convolutional networks tailored for infrared image analysis.
  • To address the issue of redundant feature extraction in homogeneous infrared image data.
  • To improve the performance of deep learning models on infrared image tasks.

Main Methods:

  • A novel pruning method based on heatmap generation metrics is introduced.
  • Network layer performance is evaluated by the number of pixel activations in generated heatmaps.
  • An Alternating training and self-pruning strategy is employed to prevent incorrect layer deletion during pruning.

Main Results:

  • The proposed heatmap-based pruning method effectively identifies and prunes redundant network layers.
  • The alternating training and self-pruning strategy reduces incorrect pruning, leading to more accurate model optimization.
  • Experimental results demonstrate performance improvements in CSPDarknet, Darknet, and Resnet models after applying the proposed method.

Conclusions:

  • The developed pruning technique enhances the efficiency and accuracy of deep convolutional networks for infrared image processing.
  • The method successfully mitigates the challenge of redundant feature extraction in homogeneous infrared data.
  • This approach offers a viable solution for optimizing deep learning models in specialized domains like infrared imaging.