Multi- stage Threshold Method for Liver Tumor Segmentation in CT scan Images and its Implementation for FPGA

Medical Imaging is a sub-field of scientific imaging. Medical image segmentation is a challenging task due to poor contrast and noise in the medical images. Medical image segmentation is applied for liver tumor detection in CT scan images in this paper. Liver tumor is pathological disorder that affects human metabolic activity. The early identification and diagnosis of liver tumor is important for curing liver cancer. In this paper a comprehensive overview of liver tumor detection using Multi stage Threshold Method is provided. It is developed using MATLAB, Xilinx System Generator (XSG) and synthesized for Spartan-6 (xc6slx16-2csg324) FPGA. The key objective is implementing liver tumor segmentationin CT scan image for an FPGA implementation with minimum device utilization and low compilation and execution time to be suitable for real-time medical applications. This paper describes in detail about hardware framework of Multi stageThreshold Method which is modeled by XSG design models (basic as well as Mcode blocks). Finally, a comparison is made in terms of power, delay and resource utilization for the implementations based on both basic and Mcode blocks. The results show reduction in resource utilized and power with slightly increase in delay using Mcode blocks compared to basic blocks..