Parallel Multithreaded Medical Images Filtering

Abstract

The quality of medical images is paramount. Being of high grade, it guarantees the quality of medical diagnosis, treatment and quality of patient’s life through the means of health care or using automate intelligent systems for medical diagnosing, treatment and monitoring. The paper presents the computational challenges in medical images processing. The great challenges are to propose parallel computational models and parallel program implementations based on the algorithms for medical images filtering. Parallel computational model based on two-dimensional filters is designed. The proposed parallel model is verified by multithreaded parallel program implementation. An investigation of the efficiency of medical images filters based on parallel multithreaded program implementation, applying two-dimensional filters on a given list of compressed jpeg medical images and generating output jpeg images for each type of applied filter. The applied filters are Brightness Control, horizontal and vertical filter of Sobel, Laplace and Blur. A number of experiments have been carried out for the case of dataset consisted of 162 whole mount slide images of Breast Cancer (BCa) specimens scanned at 40x and various number of threads. Parallel performance parameters execution time and speedup are estimated experimentally. The performance estimation and scalability analyses show that the suggested model has good scalability.