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

Coarse-to-Fine Network-Based Intra Prediction in Versatile Video Coding.

Sensors (Basel, Switzerland)·2023
See all related articles

Related Experiment Video

Updated: Oct 3, 2025

Photorealistic Learned Landscapes for Augmented Reality
06:54

Photorealistic Learned Landscapes for Augmented Reality

Published on: June 27, 2025

241

Incomplete Region Estimation and Restoration of 3D Point Cloud Human Face Datasets.

Kutub Uddin1, Tae Hyun Jeong1, Byung Tae Oh1

  • 1School of Electronics and Information Engineering, Korea Aerospace University, Goyang 10540, Korea.

Sensors (Basel, Switzerland)
|February 15, 2022
PubMed
Summary

This study introduces a novel masking method to fill missing regions in 3D point cloud human face data. The approach effectively estimates and restores incomplete 3D face scans, improving data quality for downstream tasks.

Keywords:
3D point cloudand restorationdeep learningestimationincomplete region

More Related Videos

Three-Dimensional Reconstruction of Orbital Fractures
08:18

Three-Dimensional Reconstruction of Orbital Fractures

Published on: May 16, 2025

348
Author Spotlight: Three-Dimensional Cephalometric Landmark Annotation Demonstration on Human Cone Beam Computed Tomography Scans
10:23

Author Spotlight: Three-Dimensional Cephalometric Landmark Annotation Demonstration on Human Cone Beam Computed Tomography Scans

Published on: September 8, 2023

3.1K

Related Experiment Videos

Last Updated: Oct 3, 2025

Photorealistic Learned Landscapes for Augmented Reality
06:54

Photorealistic Learned Landscapes for Augmented Reality

Published on: June 27, 2025

241
Three-Dimensional Reconstruction of Orbital Fractures
08:18

Three-Dimensional Reconstruction of Orbital Fractures

Published on: May 16, 2025

348
Author Spotlight: Three-Dimensional Cephalometric Landmark Annotation Demonstration on Human Cone Beam Computed Tomography Scans
10:23

Author Spotlight: Three-Dimensional Cephalometric Landmark Annotation Demonstration on Human Cone Beam Computed Tomography Scans

Published on: September 8, 2023

3.1K

Area of Science:

  • Computer Vision
  • 3D Data Processing
  • Machine Learning

Background:

  • Incomplete 3D point cloud data, caused by scanning issues, hinders recognition and segmentation tasks.
  • Missing data regions in 3D human face datasets significantly degrade performance.

Purpose of the Study:

  • To propose a new masking method for estimating incomplete regions in 3D point cloud human face datasets.
  • To enhance the quality of 3D face data for improved downstream applications.

Main Methods:

  • Preprocessing point cloud data, including rotation.
  • Projecting point clouds onto a 2D surface to generate masks.
  • Interpolating 2D projections and masks to estimate missing point cloud regions.
  • Utilizing a deep learning model to restore and improve the estimated point cloud quality.

Main Results:

  • The proposed method achieved competitive results on custom human face and Large-Scale Facial Model (LSFM) datasets.
  • Average Chamfer Distance (CD) of 1.30 and Hausdorff Distance (HD) of 21.46 on custom datasets.
  • Average CD of 1.35 and HD of 9.08 on LSFM datasets.
  • Demonstrated superior performance compared to existing methods.

Conclusions:

  • The masking method effectively estimates and restores incomplete 3D point cloud human face data.
  • The deep learning restoration step further enhances the quality of the estimated point clouds.
  • This approach offers a promising solution for handling missing data in 3D facial recognition and analysis.