Compressed sensing in diffuse optical tomography.

Area of Science:

• Biomedical Optics
• Medical Imaging
• Computational Imaging

Background:

• Diffuse optical tomography (DOT) enables non-invasive 3D imaging of tissue optical properties.
• Severe under-sampling in DOT often leads to image artifacts, necessitating extensive measurements.
• Traditional linear reconstruction methods may struggle with limited data acquisition.

Purpose of the Study:

• To introduce a compressed sensing (CS) framework for DOT.
• To enable improved DOT reconstructions using under-sampled data.
• To reduce the number of measurements required for accurate DOT imaging.

Main Methods:

• Implementation of a CS framework for DOT.
• Utilizing a sparsifying basis, ℓ1-regularization, and random sampling.
• Comparison with traditional singular-value decomposition (SVD) based linear reconstruction methods.

Main Results:

• CS framework demonstrated improved DOT reconstructions with under-sampled data.
• CS results showed enhanced accuracy compared to SVD with ℓ2-regularization.
• CS proved more robust against measurement reduction than traditional methods.

Conclusions:

• Compressed sensing offers a promising approach to enhance DOT imaging quality and efficiency.
• CS-DOT can achieve high-fidelity reconstructions with significantly fewer measurements.
• The CS framework provides a robust solution for under-sampled DOT challenges.