Simultaneous Activity and Attenuation Estimation in TOF-PET With TV-Constrained Nonconvex Optimization.

Area of Science:

• Medical Imaging
• Computational Science
• Optimization Theory

Background:

• Inverse problems in medical imaging often involve complex optimization.
• Simultaneous Activity and Attenuation (SAA) estimation in Time-of-Flight Positron Emission Tomography (TOF-PET) is a critical but challenging inverse problem.
• Maximum Likelihood Estimation (MLE) is a common approach for SAA, but can be computationally intensive and non-smooth.

Purpose of the Study:

• To develop a robust and efficient optimization framework for nonsmooth biconvex problems.
• To apply this framework to the SAA problem in TOF-PET using MLE.
• To introduce Total Variation (TV) constraints for improved image reconstruction.

Main Methods:

• Development of an Alternating Direction Method of Multipliers (ADMM) framework tailored for nonsmooth biconvex optimization.
• Application of the ADMM framework to MLE for SAA in TOF-PET, creating the ADMM-SAA algorithm.
• Extension of ADMM-SAA by incorporating Total Variation (TV) constraints on both activity and attenuation maps, resulting in the ADMM-TVSAA algorithm.

Main Results:

• The proposed ADMM-SAA algorithm effectively addresses the SAA problem in TOF-PET.
• The ADMM-TVSAA algorithm demonstrates improved performance by incorporating TV constraints, leading to potentially more accurate activity and attenuation maps.
• Performance comparison indicates the potential advantages of the developed algorithms over existing methods like Penalized Maximum Likelihood Activity and Attenuation Estimation (P-MLAA).

Conclusions:

• The ADMM framework provides a powerful tool for solving nonsmooth biconvex inverse problems in imaging.
• The ADMM-TVSAA algorithm offers a promising approach for accurate SAA in TOF-PET, enhancing image quality.
• This work contributes to advancing reconstruction techniques in TOF-PET imaging.