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

The Winter Nitrogen Pump: Unveiling the Accumulation and Contribution of a Long-Overlooked Nitrogen Pool.

Plant, cell & environment·2026
Same author

PIANet: a parallel interactive attention network for multi-channel PSG-based sleep staging.

Biomedical physics & engineering express·2026
Same author

Latent diffusion-based image reconstruction for near-infrared spectral tomography.

Biomedical optics express·2026
Same author

A quantitative method to compare regional tumor contrast between prone and supine breast MRI.

Frontiers in oncology·2026
Same author

Field‑Programmable Biofunctional Films: From Assisted Fabrication to Integrated Diagnostic-Therapeutic Devices.

Small (Weinheim an der Bergstrasse, Germany)·2026
Same author

Development of a high-grade glioma preclinical surgery model using an inducible KRAS/TP53 Oncopig.

Frontiers in oncology·2026

Related Experiment Video

Updated: Jun 13, 2025

Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

8.6K

Model-Based Convolution Neural Network for 3D Near-Infrared Spectral Tomography.

Chengpu Wei, Zhe Li, Ting Hu

    IEEE Transactions on Medical Imaging
    |March 3, 2025
    PubMed
    Summary

    A new Model-CNN algorithm significantly improves near-infrared spectral tomography (NIRST) image reconstruction by integrating a diffusion model with a convolutional neural network (CNN). This method enhances accuracy and speed for functional tissue imaging.

    More Related Videos

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
    04:48

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

    Published on: November 30, 2022

    2.7K
    Near Infrared Optical Projection Tomography for Assessments of β-cell Mass Distribution in Diabetes Research
    15:18

    Near Infrared Optical Projection Tomography for Assessments of β-cell Mass Distribution in Diabetes Research

    Published on: January 12, 2013

    16.4K

    Related Experiment Videos

    Last Updated: Jun 13, 2025

    Deep Learning-Based Segmentation of Cryo-Electron Tomograms
    10:25

    Deep Learning-Based Segmentation of Cryo-Electron Tomograms

    Published on: November 11, 2022

    8.6K
    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
    04:48

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

    Published on: November 30, 2022

    2.7K
    Near Infrared Optical Projection Tomography for Assessments of β-cell Mass Distribution in Diabetes Research
    15:18

    Near Infrared Optical Projection Tomography for Assessments of β-cell Mass Distribution in Diabetes Research

    Published on: January 12, 2013

    16.4K

    Area of Science:

    • Biomedical optics
    • Medical imaging
    • Computational modeling

    Background:

    • Near-infrared spectral tomography (NIRST) is a non-invasive imaging method for functional tissue assessment.
    • NIRST image reconstruction is computationally challenging due to light scattering and limited data, leading to ill-posed problems.
    • Existing methods like Tikhonov regularization and fully convolutional neural networks (FC-CNN) have limitations in accuracy and speed.

    Purpose of the Study:

    • To develop an advanced reconstruction algorithm for NIRST that overcomes the limitations of current methods.
    • To improve the accuracy and efficiency of reconstructing chromophore concentration images from NIRST data.
    • To validate the performance of the developed algorithm against established techniques using simulated, phantom, and clinical data.

    Main Methods:

    • Developed a novel reconstruction algorithm, Model-CNN, integrating a diffusion equation model with a convolutional neural network (CNN).
    • The CNN component learns a regularization prior to constrain solutions to realistic chromophore concentration distributions.
    • Evaluated Model-CNN performance by training on simulated data and testing on physical phantom and clinical NIRST datasets, comparing against Tikhonov regularization and FC-CNN.

    Main Results:

    • Model-CNN significantly outperformed Tikhonov regularization, reducing absolute bias error (ABE) by 55% for total hemoglobin (HbT) and 70% for water (H2O), and improving peak signal-to-noise ratio (PSNR) by 5.3 dB for both.
    • Processing time was reduced by 82% compared to Tikhonov regularization.
    • Model-CNN also surpassed FC-CNN, achieving 91% lower ABE for HbT and 75% for H2O, with PSNR improvements of 7.3 dB and 4.7 dB, respectively, despite being trained only on simulation data.

    Conclusions:

    • The Model-CNN algorithm offers a superior approach for NIRST image reconstruction, providing enhanced accuracy and efficiency.
    • This method effectively addresses the ill-posed nature of NIRST reconstruction, even when trained on simplified simulation data.
    • Model-CNN demonstrates strong potential for clinical application in functional tissue imaging, offering improved diagnostic capabilities.