parallel programming application on medical image processing: mri contours matching algorithm based on gpu accelerated methods for tumor differential analysis

the goal of this paper is to investigate new methods of tumor progress study using machine vision, image processing and parallel programming techniques and proposing a solution to boost performance of magnetic resonance images differential detection that supports large scale medical image processing systems in determining the activity of tumors. after detection the tumor, we need to watch it out and log any activity and measuring rate of the tumor, growth, scattering, even decline and eliminating progresses. following method can make it possible to compress the records of patient's mr or ct scan images and help physician and medics to scan every progress of tumor or even septicity's area, thus detect the result of efficiency of treatment. now by use of parallel programming and using modern graphic processing units we're able to do this processing steps rapidly and apply this method on huge amounts of medical images, network based or by a personal computer.