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 Experiment Video

Updated: Dec 27, 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

938

Training Convolutional Neural Networks withMulti-Size Images and Triplet Loss for RemoteSensing Scene Classification.

Jianming Zhang1, Chaoquan Lu1, Jin Wang1,2

  • 1Hunan Provincial Key Laboratory of Intelligent Processing of Big Data on Transportation, School ofComputer and Communication Engineering, Changsha University of Science and Technology, Changsha 410114, China.

Sensors (Basel, Switzerland)
|February 27, 2020
PubMed
Summary

Related Concept Videos

Force Classification01:22

Force Classification

2.2K
Forces play a crucial role in the study of physics and engineering. They are essential in describing the motion, behavior, and equilibrium of objects in the physical world. Forces can be classified based on their origin, type, and direction of action.
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...
2.2K

You might also read

Related Articles

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

Sort by
Same author

Dynamic functional network connectivity analysis in schizophrenia based on a spatiotemporal CPD framework.

Journal of neural engineering·2024
Same author

Unmanned Aerial Vehicle Assisted Post-Disaster Communication Coverage Optimization Based on Internet of Things Big Data Analysis.

Sensors (Basel, Switzerland)·2023
Same author

A Microservice and Serverless Architecture for Secure IoT System.

Sensors (Basel, Switzerland)·2023
Same author

<i>Game-o-Meta</i>: Trusted Federated Learning Scheme for P2P Gaming Metaverse beyond 5G Networks.

Sensors (Basel, Switzerland)·2023
Same author

UsCoTc: Improved Collaborative Filtering (CFL) recommendation methodology using user confidence, time context with impact factors for performance enhancement.

PloS one·2023
Same author

Innovative Spectrum Handoff Process Using a Machine Learning-Based Metaheuristic Algorithm.

Sensors (Basel, Switzerland)·2023
This summary is machine-generated.

This study enhances remote sensing scene classification accuracy by incorporating triplet loss, dropout, and multi-size image training during model development. These methods improve performance without increasing computational costs during inference.

Area of Science:

  • Computer Science
  • Remote Sensing
  • Machine Learning

Background:

  • Remote sensing scene classification algorithms often rely on additional modules, increasing computational overhead.
  • There is a need for methods to improve classification accuracy without augmenting model complexity at inference.

Purpose of the Study:

  • To enhance remote sensing scene classification accuracy.
  • To achieve this without increasing model parameters or computational overhead during inference.

Main Methods:

  • Implemented a network training strategy using multi-size images.
  • Introduced triplet loss with a dedicated branch for enhanced supervision.
  • Incorporated dropout between the feature extractor and classifier to prevent overfitting.
Keywords:
dropoutremote sensing scene classificationtraining with multi‐size imagestriplet loss

Related Experiment Videos

Last Updated: Dec 27, 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

938

Main Results:

  • The combined approach improved overall accuracy by 1.61% on the OPTIMAL dataset.
  • Ablation studies showed individual gains of 0.53% (dropout), 0.38% (triplet loss), and 0.7% (multi-size images).
  • The model demonstrated competitive performance against existing algorithms on AID, NWPU-RESISC45, and OPTIMAL datasets.

Conclusions:

  • The proposed training strategies (multi-size images, triplet loss, dropout) effectively boost remote sensing scene classification accuracy.
  • These improvements are achieved without increasing model parameters or computational load during inference.
  • Multi-size image training yielded the most significant accuracy gain individually, but the combination of all three methods proved most effective.