Iterative reconstruction of CT images on GPUs

Area of Science:

• Medical Imaging
• Computational Imaging
• Nuclear Medicine

Background:

• Iterative reconstruction is widely used in nuclear medicine imaging (gamma-camera, SPECT, PET) but not yet in CT.
• Large datasets in CT make iterative reconstruction computationally intensive.
• Graphical Processing Units (GPUs) offer a potential solution to mitigate the high computational cost.

Purpose of the Study:

• To develop a GPU-based algorithm for iterative reconstruction in CT.
• To reconstruct high-quality images from under-sampled and noisy projection data.

Main Methods:

• Implementation of an iterative reconstruction algorithm on Graphical Processing Units (GPUs).
• Processing of under-sampled and noisy projection data specific to CT imaging.

Main Results:

• Demonstration of a GPU-based approach for iterative reconstruction in CT.
• Successful reconstruction of high-quality images despite data limitations.

Conclusions:

• GPU acceleration can overcome the computational challenges of iterative reconstruction in CT.
• This method enables high-quality image reconstruction from under-sampled and noisy CT data.