Sparsity-promoting fluorescence molecular tomography with iteratively reweighted regularization.

Area of Science:

• Biomedical imaging
• Optical imaging
• Molecular imaging

Background:

• Fluorescence molecular tomography (FMT) is a key technique for in vivo small animal imaging.
• The ill-posed nature of FMT often necessitates Tikhonov regularization, which can lead to over-smoothed images.
• Accurate source localization is crucial for interpreting FMT data.

Purpose of the Study:

• To develop a more accurate method for fluorescence molecular tomography reconstruction.
• To address the over-smoothing issue associated with Tikhonov regularization in FMT.
• To leverage the sparsity of fluorescent sources for improved image reconstruction.

Main Methods:

• Implemented an iteratively reweighted scheme to approximate L0- or L1-norm regularization.
• Utilized the sparsity of the fluorescent source as a priori information.
• Dynamically updated the weight matrix during the iterative process.

Main Results:

• The proposed method successfully preserved the sparsity of the fluorescent source.
• Effective reconstruction was achieved even with limited measurement data.
• The new approach offers improved spatial resolution compared to traditional methods.

Conclusions:

• The iteratively reweighted scheme provides a robust alternative to Tikhonov regularization for FMT.
• This method enhances the accuracy of in vivo small animal imaging by promoting sparse solutions.
• The findings have significant implications for preclinical research and molecular imaging applications.