Detail publikace

Using a GPU to Accelerate Electrical Impedance Tomography Algorithms

Originální název

Using a GPU to Accelerate Electrical Impedance Tomography Algorithms

Anglický název

Using a GPU to Accelerate Electrical Impedance Tomography Algorithms

Jazyk

en

Originální abstrakt

This paper discusses parallelizing an algorithm for image reconstruction in EIT. A GPU NVIDIA platform with CUDA kernel was used for data processing. Introductory section comprises description of EIT and relevant practical applications. The following chapter outlines objective function with Tikhonov regularization and possibility of mathematical extension to increase image resolution. Besides we also characterize the modification of reconstruction algorithm with respect to computional speed. The achieved results are discussed with a survey of the achieved algorithm acceleration and related problems.

Anglický abstrakt

This paper discusses parallelizing an algorithm for image reconstruction in EIT. A GPU NVIDIA platform with CUDA kernel was used for data processing. Introductory section comprises description of EIT and relevant practical applications. The following chapter outlines objective function with Tikhonov regularization and possibility of mathematical extension to increase image resolution. Besides we also characterize the modification of reconstruction algorithm with respect to computional speed. The achieved results are discussed with a survey of the achieved algorithm acceleration and related problems.

BibTex


@inproceedings{BUT139897,
  author="Jan {Dušek} and Jan {Mikulka}",
  title="Using a GPU to Accelerate Electrical Impedance Tomography Algorithms",
  annote="This paper discusses parallelizing an algorithm for image reconstruction in EIT. A GPU NVIDIA platform with CUDA kernel was used for data processing. Introductory section comprises description of EIT and relevant practical applications. The following chapter outlines objective function with Tikhonov regularization and possibility of mathematical extension to increase image resolution. Besides we also characterize the modification of reconstruction algorithm with respect to computional speed. The achieved results are discussed with a survey of the achieved algorithm acceleration and related problems.",
  booktitle="Proceedings of IIPhDW 2017 in Lodz",
  chapter="139897",
  howpublished="online",
  year="2017",
  month="september",
  pages="59--63",
  type="conference paper"
}