Review on FPGA-based Hardware Accelerators of CNN for Healthcare Applications

Preview this chapter:

Review on FPGA-based Hardware Accelerators of CNN for Healthcare Applications, Page 1 of 1

< Previous page | Next page > /docserver/preview/fulltext/9789815274134/chapter-4-1.gif

In recent years, Deep Convolutional Neural Network (CNN) has been the fastest-growing area of Artificial Neural Network (ANN). In addition to image classification and segmentation, CNN can detect objects in video and recognize speech. This is because CNNs take a lot of computation. The CNN function also lends itself to programmable hardware such as Field Programmable Gate Arrays (FPGAs). Recently, hardware accelerators have become incredibly popular for a broad spectrum of healthcare applications. The emergence of edge computing has made it possible to combine a large number of sensors and process information using lightweight computing. Deep learning algorithms have advanced significantly over time, providing intriguing prospects for their use even in safety-sensitive biomedical and healthcare applications. This study presents a thorough analysis and discussion of several difficulties in the implementation of FPGA-based hardware acceleration for healthcare applications. There are some clear advantages that a variety of generalized new architectures and devices have over traditional processing units. This survey is expected to be useful for researchers in the area of artificial intelligence, FPGA-based hardware accelerators of CNN for Biomedical applications, and system design.