Related Experiment Videos
Three-dimensional, Bayesian image reconstruction from sparse and noisy data sets: near-infrared fluorescence
Margaret J Eppstein1, Daniel J Hawrysz, Anuradha Godavarty
1Department of Computer Science, 327 Votey Building, University of Vermont, Burlington, VT 05405, USA. Maggie.Epstein@uvm.edu
Summary
This study demonstrates a Bayesian reconstruction method for near-infrared (NIR) fluorescence imaging. The technique accurately maps optical properties in large tissues, even with noisy, sparse data, improving diagnostic imaging quality.
Area of Science:
- Biomedical Optics
- Medical Imaging
- Fluorescence Imaging
Background:
- Near-infrared (NIR) fluorescence imaging offers potential for deep tissue visualization.
- Challenges exist in reconstructing accurate optical property maps from sparse surface measurements, especially in large volumes with variable signal-to-noise ratios (SNR).
Purpose of the Study:
- To demonstrate a robust method for inverting sparse NIR fluorescence measurements to recover interior optical property maps.
- To address the spatial variability in SNR inherent in large-volume medical imaging applications.
Main Methods:
- Utilized a Bayesian minimum-variance reconstruction algorithm for image recovery.
- Employed frequency-domain photon migration measurements on tissue-simulating phantoms with embedded heterogeneities.
- Incorporated experimentally determined measurement error variance, recursively updated parameter uncertainty, and dynamic zonation for improved reconstruction.
Main Results:
- Successfully reconstructed spatial absorption estimates from limited surface measurements (160-296) on large phantoms.
- Demonstrated high variability in measurement error at fluorescent emission wavelengths.
- Showcased improved convergence and reconstruction quality through Bayesian conditioning.
Conclusions:
- Bayesian reconstruction effectively compensates for spatial SNR variability in NIR fluorescence imaging.
- Accurate optical property mapping in large volumes requires reconstruction methods that account for measurement SNR range.
- The developed method shows promise for NIR contrast-enhanced diagnostic medical imaging.