Model-Based Preprocessing FPGA Practice for DSA Images

This presentation describes how Siemens Healthineers uses model-based design to improve FPGA development efficiency for a vascular controller (DSA), a key device for intraoperative real-time navigation. After building a pixel flow algorithm model, high-performance HDL code is automatically generated. This frees developers from the tedious tasks of HDL design and timing debugging, allowing them to focus more on algorithm-level optimization.

The presentation details Siemens Healthineers' model-based approach to vascular controller development, from building the medical image pre-processing algorithm model, data type conversion, HDL code generation, automated FPGA resource and timing optimization, and ultimately hardware-in-the-loop testing. This significantly improves development efficiency while ensuring key performance indicators such as FPGA design timing.