Noise Removal in Medical Images Using Pulse Coupled Neural Networks

作者：Nisha S Shajun^*; Mohideen S Kother

来源：Journal of Medical Imaging and Health Informatics, 2017, 7(1): 101-105.

DOI：10.1166/jmihi.2017.1990

• Medical Images
• Noise Removal
• Pulse Coupled Neural Networks
• Filtering Techniques

In this paper, we propose a new noise removal technique for medical images using a Pulse-Coupled Neural Network (PCNN) combined with traditional filtering methods. We consider four different variants such as the neuromirne structure, intersecting cortical model, unit-linking model and multi-channel model. The Wiener, median, average and Gaussian filters are used in conjunction with the PCNN for effective noise removal. A mixture of the speckle and Gaussian are considered for a CT skull image, while a mixture of the Rician and Gaussian are considered for an MRI brain image, and a mixture of the Gaussian and salt-and-pepper are considered for a mammogram image. Performance metrics such as the peak signal-to-noise ratio, weighted signal-to-noise ratio, visual signal-to-noise ratio and the structural similarity index are used to evaluate the performance of noise removal. From the results, it is observed that the multichannel model in combination with the median filter performs best for noise removal in medical images.

• 出版日期2017-2