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

You might also read

Related Articles

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

Sort by
Same author

Integration of improved YOLOv5 for face mask detector and auto-labeling to generate dataset for fighting against COVID-19.

The Journal of supercomputing·2023
Same author

Comparison of Pre-Trained YOLO Models on Steel Surface Defects Detector Based on Transfer Learning with GPU-Based Embedded Devices.

Sensors (Basel, Switzerland)·2022
Same author

Block-HPCT: Blockchain Enabled Digital Health Passports and Contact Tracing of Infectious Diseases like COVID-19.

Sensors (Basel, Switzerland)·2022
Same author

A Deep Learning Method for 3D Object Classification and Retrieval Using the Global Point Signature Plus and Deep Wide Residual Network.

Sensors (Basel, Switzerland)·2021
Same author

A Classification Method for the Cellular Images Based on Active Learning and Cross-Modal Transfer Learning.

Sensors (Basel, Switzerland)·2021
Same author

Consistency of Medical Data Using Intelligent Neuron Faster R-CNN Algorithm for Smart Health Care Application.

Healthcare (Basel, Switzerland)·2020

Related Experiment Video

Updated: Jan 10, 2026

Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures
09:10

Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures

Published on: August 5, 2021

2.2K

Efficient Synthetic Defect on 3D Object Reconstruction and Generation Pipeline for Digital Twins Smart Factory.

Viet-Hoan Nguyen1,2, Thi-Ngot Pham3,4, Jun-Ho Huh3,4

  • 1R&D Center, Intown Co., Ltd., Busan 08592, Republic of Korea.

Sensors (Basel, Switzerland)
|November 27, 2025
PubMed
Summary

This study introduces a framework for 3D industrial object reconstruction and synthetic defect generation, enhancing digital twins for smart factories. The synthetic data significantly improved object detection models, demonstrating its value in computer vision applications.

Keywords:
3D model reconstructionNERFYOLOobject detectionsynthetic data generation

More Related Videos

Simulation of a Scaled Assembly Process with Collaboration of a Robotic Arm and Monitoring through a Vision System for Quality Control
05:47

Simulation of a Scaled Assembly Process with Collaboration of a Robotic Arm and Monitoring through a Vision System for Quality Control

Published on: August 29, 2025

406
Three-Dimensional Reconstruction of Orbital Fractures
08:18

Three-Dimensional Reconstruction of Orbital Fractures

Published on: May 16, 2025

597

Related Experiment Videos

Last Updated: Jan 10, 2026

Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures
09:10

Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures

Published on: August 5, 2021

2.2K
Simulation of a Scaled Assembly Process with Collaboration of a Robotic Arm and Monitoring through a Vision System for Quality Control
05:47

Simulation of a Scaled Assembly Process with Collaboration of a Robotic Arm and Monitoring through a Vision System for Quality Control

Published on: August 29, 2025

406
Three-Dimensional Reconstruction of Orbital Fractures
08:18

Three-Dimensional Reconstruction of Orbital Fractures

Published on: May 16, 2025

597

Area of Science:

  • Computer Vision
  • Digital Twins
  • 3D Reconstruction

Background:

  • High-quality 3D objects are vital for digital twins.
  • Synthetic data is crucial for deep learning in computer vision.
  • Real-world defect data collection is challenging and costly.

Purpose of the Study:

  • To develop a unified framework for 3D industrial object reconstruction and synthetic defect generation.
  • To improve the realism of synthetic data for industrial applications.
  • To enhance object detection performance using synthetic defect data.

Main Methods:

  • Utilized Neural Radiance Fields (NeRF) for 3D object reconstruction from smartphone videos.
  • Compared NeRF variants (Instant-NGP, Nerfacto, Volinga, Tensorf) within the Nerfstudio framework.
  • Generated synthetic defects on refined 3D models using NVIDIA Omniverse Replicator.
  • Evaluated object detection using YOLO models with synthetic and real-plus-synthetic defect datasets.

Main Results:

  • Instant-NGP and Nerfacto demonstrated superior performance in 3D object reconstruction.
  • Synthetic defect data improved the generalization capability of YOLO models.
  • mAP@0.5 accuracy enhancement ranged from 1.5% to 18.8% across different YOLO versions.

Conclusions:

  • The proposed framework effectively generates realistic synthetic defect data for industrial applications.
  • Synthetic data significantly boosts the performance and generalization of object detection models.
  • This approach offers a viable solution for data scarcity in industrial computer vision.