Graphics processing unit accelerating compressed sensing photoacoustic computed tomography with total variation.

Area of Science:

• Biomedical Imaging
• Computational Imaging
• Medical Physics

Background:

• Compressed sensing photoacoustic computed tomography (CS-PACT) is vital for sparse-sampling PACT.
• The iterative image reconstruction process in CS-PACT is computationally intensive and time-consuming.

Purpose of the Study:

• To develop a graphics processing unit (GPU)-based parallel computation framework for accelerating total-variation-based CS-PACT.
• To adapt and validate this framework on a custom-made PACT system.

Main Methods:

• Extracted five compute-intensive operators from the iterative algorithm.
• Redesigned these operators for parallel processing on a GPU.
• Implemented and tested the framework on a custom PACT system.

Main Results:

• Achieved image reconstruction speeds 24-31 times faster compared to CPU performance.
• Demonstrated the feasibility of the GPU-accelerated CS-PACT method through in vivo experiments on human hands.

Conclusions:

• The developed GPU-based parallel computation framework significantly enhances the speed of CS-PACT image reconstruction.
• This acceleration makes CS-PACT a more practical and efficient imaging strategy, particularly for sparse-sampling applications.