Xilinx technology offers a heterogeneous and a highly distributed architecture, delivering more compute for image processing, enabling healthcare companies to offer effective and faster diagnosis. The highly integrated Xilinx Zynq UltraScale+ MPSoC with its adaptable Field Programmable Gate Arrays (FPGA), integrated accelerators for deep learning (DPU), integrated digital signal processor (DSP) and its ARM® multi-processor systems perform accurate image classification and detection with AI inferencing in close to real-time with low latency and low power consumption.
PYNQ™, the open-source Python programming platform for Zynq architecture with Xilinx’s latest AI Toolkit Vitis AI™ version 1.1, is used to compile the deep learning models for running accelerated inference making this solution cost-effective. Xilinx unified software platforms, Vitis for application development and Vitis AI for optimizing and deploying accelerated ML inference, mean that data scientists and algorithm developers can use advanced devices like the MPSoC easily in their projects.
New healthcare workflows need to deliver more compute for image processing, data privacy, security, patient safe ty and accuracy in much smaller edge devices. Heterogeneous and adaptable distributed systems, which can be small and portable, are key for solving this problem. Xilinx devices like the Zynq UltraScale+ MPSoC and the Vitis software platform are ideal for delivering the optimized clinical device enabled for AI inferencing at the edge.