Multilevel, hybrid regularization method for reconstruction of fluorescent molecular tomography

Area of Science:

• Biomedical Imaging
• Computational Science
• Optical Engineering

Background:

• Fluorescent Molecular Tomography (FMT) is an ill-posed inverse problem.
• Accurate reconstruction in FMT is challenging due to limited data and inherent noise.
• Existing regularization methods often struggle with stability and resolution.

Purpose of the Study:

• To develop a robust and stable multilevel, hybrid regularization method for FMT.
• To improve the accuracy and feasibility of FMT reconstruction.
• To leverage the sparse nature of fluorescent probes in FMT.

Main Methods:

• A hybrid regularization approach combining sparse and Landweber iterative regularization.
• Utilizing the hp-finite element method (hp-FEM) for a multilevel reconstruction strategy.
• Applying sparse regularization on coarse mesh levels and Landweber on refined levels.

Main Results:

• Demonstrated improved stability in solving the ill-posed inverse problem of FMT.
• Achieved accurate reconstructions in both simulated (3D digital mouse atlas) and physical (phantom) experiments.
• Validated the potential and feasibility of the proposed hybrid regularization method.

Conclusions:

• The proposed multilevel, hybrid regularization method offers a significant advancement for FMT.
• This technique enhances the reliability and accuracy of molecular imaging reconstructions.
• The method shows promise for various biomedical applications requiring precise FMT imaging.