An intelligent system for segmenting an abdominal image in multi core architecture

Image Segmentation is the key process in medical imaging which produces different ROI (Region of Interests) as per needed by the medical practitioner, who uses it for further treatment planning. Fast processing of the medical image is a major requirement in today's scenario. In the case of abdominal CT/MR Scans images due to the dimension and the presence of multiple regions like liver, kidney, spleen etc fast processing is very much required. This paper proposed a method for kidney segmentation from an abdominal image and it suggest the idea about segmentation of multiple regions like Spine, Kidney, Liver of Abdominal image in order to produce efficient and prompt result. All the activated cores of a particular processor are responsible to handle their individual region. The main approach of the proposed system is the allocation of spine as a land mark for coordinate reference in a particular core. This is more reliable to identify kidney region seed point identification while moving towards on the axis of centroid of spine. Kidneys region can be easily identify on different cores for faster execution. A new core can be activated for liver segmentation for this we have used prior knowledge of the image. This proposed method is the efficient way to assign particular abdominal region to particular core in available Multicore Architecture to execute speedy and efficient result rather than the existing parallel implementation. Proposed method demonstrate that the parallel implementation of region wise segmentation of abdominal CT/MR image results in a very good speed-up compared with sequential implementation on a core i3 processor and Intel Xeon Processor.