Implementation of Medical Image Processing Algorithms on FPGA Using Xilinx System Generator

While software implementation of various image processing methods is adequate for general application, when it comes to meeting real‐time requirements, the implementation has to be performed in hardware. In applications like digital signals and handling massive data in particular in real time, field programmable gate arrays (FPGAs) have many advantages. FPGAs have now become a part of digital signal processing (DSP), mostly due to the ever‐decreasing cost and reconfigurability. Xilinx System Generator (XSG) is a Xilinx tool that expands Simulink models to allow for FPGA design to be built inside MATLAB. This chapter deals with implementation of different image processing algorithms on medical images using XSG for the purposes of translating them to hardware. Several algorithms such as contrast stretching and edge detection (Robert, Prewitt, Sobel, and Canny) are described in detail and compared. Results indicate that Prewitt and Sobel detectors achieve better results than Robert, while Canny method outputs thinner edges. However, the Robert operator is the least resource demanding in comparison to other methods. These implemented algorithms can be extended and used as a part of more complex designs in medical image processing on FPGA.