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

Ultrafast carrier dynamics and electronic properties of PtSe<sub>2</sub>/MoSe<sub>2</sub> and WSe<sub>2</sub> 2D TMDC layered structures on mica: combined THz spectroscopy and DFT study.

RSC advances·2026
Same author

Development of Polarization-Insensitive THz-to-IR Converters for Low-IR-Signature Target Detection and Imaging.

Sensors (Basel, Switzerland)·2024
Same author

Emotion Classification Based on Pulsatile Images Extracted from Short Facial Videos via Deep Learning.

Sensors (Basel, Switzerland)·2024
Same author

Atmospheric Turbulence Degraded Video Restoration with Recurrent GAN (ATVR-GAN).

Sensors (Basel, Switzerland)·2023
Same author

3D Object Detection via 2D Segmentation-Based Computational Integral Imaging Applied to a Real Video.

Sensors (Basel, Switzerland)·2023
Same author

Fast and Enhanced MMW Imaging System Using a Simple Row Detector Circuit with GDDs as Sensor Elements and an FFT-Based Signal Acquisition System.

Sensors (Basel, Switzerland)·2023

Related Experiment Video

Updated: Feb 28, 2026

Holistic Facial Composite Creation and Subsequent Video Line-up Eyewitness Identification Paradigm
09:49

Holistic Facial Composite Creation and Subsequent Video Line-up Eyewitness Identification Paradigm

Published on: December 24, 2015

14.6K

Concealed Face Analysis and Facial Reconstruction via a Multi-Task Approach and Cross-Modal Distillation in Terahertz

Noam Bergman1, Ihsan Ozan Yildirim2, Asaf Behzat Sahin3

  • 1Department of Electro-Optical Engineering, School of Electrical and Computer Engineering, Ben-Gurion University of the Negev, 1 Ben-Gurion Blvd, Beer Sheva 8410501, Israel.

Sensors (Basel, Switzerland)
|February 27, 2026
PubMed
Summary
This summary is machine-generated.

This study introduces a novel Multi-Task Learning network for Terahertz (THz) imaging to improve concealed face recognition and reconstruction. A cross-modal approach enhances THz-only biometrics, overcoming sparsity and noise limitations.

Keywords:
THz facial reconstructioncross-modal fusiondeep learningfacial biometricsknowledge distillationmulti-task learningterahertz imaging

More Related Videos

Author Spotlight: Revolutionizing Remote Surgery with Augmented Reality and Robotics for Enhanced Precision and Accessibility
07:46

Author Spotlight: Revolutionizing Remote Surgery with Augmented Reality and Robotics for Enhanced Precision and Accessibility

Published on: August 9, 2024

1.2K
High-resolution, High-speed, Three-dimensional Video Imaging with Digital Fringe Projection Techniques
11:34

High-resolution, High-speed, Three-dimensional Video Imaging with Digital Fringe Projection Techniques

Published on: December 3, 2013

16.1K

Related Experiment Videos

Last Updated: Feb 28, 2026

Holistic Facial Composite Creation and Subsequent Video Line-up Eyewitness Identification Paradigm
09:49

Holistic Facial Composite Creation and Subsequent Video Line-up Eyewitness Identification Paradigm

Published on: December 24, 2015

14.6K
Author Spotlight: Revolutionizing Remote Surgery with Augmented Reality and Robotics for Enhanced Precision and Accessibility
07:46

Author Spotlight: Revolutionizing Remote Surgery with Augmented Reality and Robotics for Enhanced Precision and Accessibility

Published on: August 9, 2024

1.2K
High-resolution, High-speed, Three-dimensional Video Imaging with Digital Fringe Projection Techniques
11:34

High-resolution, High-speed, Three-dimensional Video Imaging with Digital Fringe Projection Techniques

Published on: December 3, 2013

16.1K

Area of Science:

  • Biometrics and Imaging Technologies
  • Artificial Intelligence and Machine Learning
  • Signal Processing

Background:

  • Terahertz (THz) imaging enables stand-off biometrics but faces challenges like sparsity, low resolution, and high noise.
  • Existing THz biometric systems struggle with data limitations and image quality.

Purpose of the Study:

  • To develop a unified Multi-Task Learning (MTL) network for enhanced Terahertz (THz) facial biometrics.
  • To improve concealed face verification, facial posture classification, and generative reconstruction using THz data.
  • To explore cross-modal knowledge distillation to enhance THz-only student models.

Main Methods:

  • A novel U-Net-like encoder-based Multi-Task Learning (MTL) network was designed for concealed THz facial data.
  • The network simultaneously addressed face verification, posture classification, and generative reconstruction tasks.
  • A cross-modal teacher-student approach was employed, using visible-spectrum data to guide a THz-only student model during distillation.

Main Results:

  • The unified MTL network demonstrated highly successful performance on a challenging dataset of THz facial images.
  • The cross-modal distilled student model showed improved latent space separability compared to the single-modality baseline.
  • Both THz-only and distilled models maintained high fidelity in generating unconcealed faces from concealed inputs.

Conclusions:

  • The proposed MTL network effectively addresses limitations in THz facial biometrics.
  • Cross-modal distillation offers a viable strategy to enhance THz-only biometric systems.
  • The developed models show promise for robust concealed biometrics using Terahertz imaging.