Arm Processor based CT-MRI fusion using PYNQ Z2 for Brain Tumor Diagnosis

Several system architectures are needed for the computationally demanding process of image fusion in order to get high resolution images as well as real-time apps. The Zynq SoC, integrated with processor and Field Programmable Gate Array (FPGA), can address these issues. The High-Level Synthesis (HLS) tools shorten FPGA implantation time, while Python Productivity for Zynq (PYNQ) framework boosts productivity. We used the multiscale transformation techniques to demonstrate image fusion of Computed Tomography (CT) and Magnetic Resonance imaging (MRI) on Zynq Multiprocessor System on-chip (MPSoC) device. We illustrated how the PYNQ framework and HLS tools allow for the rapid deployment of CT-MRI fusion, and how FPGAs may speed up the fusion of CT-MRI images. The suggested method is compared to two other existing CT-MRI image fusion methods, NSCT-PA-PCNN and NSCT-PCNN. Five objective assessment metrics such as mutual information (MI), structural similarity index measure (SSIM), weighted edge information (Q<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AB/F</sup>), visual information fidelity factor (VIFF), and average gradient (AG) were used for evaluation of proposed method and existing methods. According to the experimental results, our suggested approach performed better than the current approaches in terms of efficiently preserving the edges and extracting information from the source images. MI has increased by 18% in the case of acute stroke disease and by 2% in the case of cerebral toxoplasmosis disease with reference to NSCT-PCNN.