Related Experiment Video
Updated: Jul 17, 2026
08:52
Software-Assisted Quantitative Measurement of Osteoarthritic Subchondral Bone Thickness
Published on: March 18, 2022
Chromatic differential confocal matrix-based 3D topography and hyperspectral imaging with deep learning for
Xiaer Zou1,2, Jiajing Ye2,3, Dawei Gong2
1Centre for Optical and Electromagnetic Research, College of Optical Science and Engineering, Zhejiang University, Hangzhou 310058, China.
Biomedical Optics Express
|July 16, 2026
Summary
This study introduces a novel 3D spatio-spectral framework for knee osteoarthritis grading, enhancing diagnostic accuracy. The deep learning approach combines 3D topography and hyperspectral imaging for objective assessment.
Area of Science:
- Biomedical Engineering
- Medical Imaging
- Artificial Intelligence
Background:
- Current arthroscopy for knee osteoarthritis (KOA) relies on subjective visual inspection, limiting accurate assessment.
- Objective and quantitative methods are needed to improve KOA diagnosis and grading.
Purpose of the Study:
- To develop and evaluate a 3D spatio-spectral automatic grading framework for knee osteoarthritis.
- To integrate deep learning with advanced imaging techniques for enhanced KOA assessment.
Main Methods:
- A novel framework combining chromatic differential confocal matrix-based 3D topography and hyperspectral imaging (CDCM-THI) was developed.
- A digital micromirror device captured high-resolution 3D micro-topography and hyperspectral cartilage signatures.
- An improved ResNet50 late fusion network with a consistency-enhanced loss function was used for data analysis.
Main Results:
- The framework achieved an average cross-validation accuracy of 94.83% on ex vivo clinical samples.
- The proposed spatio-spectral approach significantly outperformed single-modal imaging baselines.
- The system simultaneously captured detailed topographical and spectral information of cartilage.
Conclusions:
- The developed non-destructive framework offers a powerful quantitative tool for KOA assessment.
- This approach bridges topographical and optical analysis for improved diagnostic capabilities.
- The framework has potential for integration into future deep learning-assisted arthroscopic procedures.
