Voxel-based 3D MRI volume segmentation system using real-time hardware architecture

In this paper, a hard real time and optimized multi-level 3D MRI volume voxel-based Fractional Order Darwinian Particle Swarm Optimization (FODPSO) Segmentation algorithm system has been developed. It is an improved hardware design based on Fractional Order PSO to ensure non-convergence to an erroneous desired threshold. The Darwinian variant is used to guarantee a more faithful multi-level segmentation (3-levels for this work). The hardware design was proposed based on the Vivado System Generator (VSG) tool. Co-simulation and experimental results validate the performance of the proposed system based on the BraTS 2013 dataset. Also, resource utilization, execution time, power consumption, efficiency, specificity, and sensitivity prove the quality of the proposed system in terms of architecture and in terms of the performance of the multi-level segmentation of voxel-based 3D MRI volumes.