Speaker
Description
Artificial intelligence (Deep Learning) is everywhere. The space industry is no exception. Automated recognition of lunar craters for moon landings and identification of space junk using imaging could play important roles in securing space safety and advancing space exploration. Deep Learning is the most successful solution for image-based object classification, and for most practical applications it requires performant platforms like FPGAs and SoCs.
Designing Deep Learning networks for embedded devices such as FPGAs and SoCs is challenging because of resource constraints, the complexity of programming in Verilog or VHDL, and the hardware expertise needed for prototyping on an FPGA or SoC.
In this talk I will explain how to prototype and deploy Deep Learning-based vision applications on FPGAs and SoCs using MATLAB. Starting with a pretrained model, I will demonstrate a MATLAB-based workflow that deploys the trained network for image recognition to the Xilinx Ultrascale+ MPSoC platform for inference using APIs from MATLAB.
Deep Learning practitioners can quickly explore different networks and evaluate their performance on FPGAs or SoCs directly from MATLAB. This workflow also enables hardware engineers to optimize and generate portable Verilog and VHDL code that can be integrated with the rest of their applications.
