Speed-up of the medical image processing using the parallel FPGA system

摘要

Medical image processing often needs tremendous processing power, thus general-purpose computers consume much time to process. Dedicated LSIs can efficiently execute the application, however, they require much time and costs to develop and they are less flexible for variants of applications. For such applications, Field-Programmable Gate Arrays (FPGAs) are suitable. In this paper, we propose a parallel FPGA system applicable to automatic detection of stomach cancer from a medical film image. The system accelerates heavy calculation of concentration indices. According to our preliminary evaluation by using hardware description language, the system achieves at most 47.72 times faster than ordinal workstations.