Use of Split Bregman denoising for iterative reconstruction in fluorescence diffuse optical tomography

Area of Science:

• Biomedical Imaging
• Optical Techniques
• Medical Physics

Background:

• Fluorescence diffuse optical tomography (fDOT) enables noninvasive quantification of fluorescent tracers in small animals.
• Iterative methods, like the algebraic reconstruction technique (ART), are standard for fDOT image reconstruction.
• Advanced techniques like l1-regularization and the Split Bregman (SB) method show promise for signal recovery and image reconstruction.

Purpose of the Study:

• To introduce a novel approach, ART-SB, for addressing the fDOT inverse problem.
• To enhance the accuracy and efficiency of fDOT image reconstruction.

Main Methods:

• The ART-SB method integrates cost-efficient ART iterations with an SB-based denoising step.
• The SB method efficiently solves the total variation minimization for image denoising.
• The approach was validated using both simulated and experimental fDOT data.

Main Results:

• ART-SB demonstrated substantial benefits over the conventional ART method.
• The combined approach offers improved image reconstruction quality in fDOT.

Conclusions:

• The proposed ART-SB method represents a significant advancement in fDOT image reconstruction.
• ART-SB provides a powerful tool for quantitative molecular imaging in small animal research.