Accelerated organ region segmentation by the revised radial basis function network using a graphics processing unit.

Area of Science:

• Medical Imaging
• Computational Biology
• Computer Science

Background:

• Medical image segmentation is crucial for diagnosis and treatment planning.
• Traditional segmentation methods can be time-consuming, limiting clinical applications.
• The revised radial basis function (RBF) network offers a potential solution for accurate segmentation.

Purpose of the Study:

• To accelerate organ segmentation using a revised RBF network.
• To evaluate the efficiency of graphics processing unit (GPU) acceleration compared to central processing unit (CPU) processing.
• To assess the impact of GPU acceleration on segmentation accuracy.

Main Methods:

• Segmentation of lung and liver regions from CT images using a revised RBF network.
• Comparison of processing times: serial CPU, parallel CPU (4 cores), and GPU.
• Evaluation of segmentation accuracy using concordance rates.

Main Results:

• GPU processing significantly reduced segmentation times: lung (20.16s) and liver (11.02s) compared to serial CPU (211.03s, 124.21s) and parallel CPU (57.80s, 35.35s).
• High segmentation accuracy was maintained: 98% for lung and 96% for liver.
• GPU acceleration reduced segmentation time by over 90%.

Conclusions:

• GPU acceleration of the revised RBF network effectively reduces organ segmentation time in medical imaging.
• This acceleration does not compromise segmentation accuracy.
• The GPU-accelerated method offers a valuable tool for reducing workload in medical image analysis.