eduzhai > Applied Sciences > Engineering >

Hardware Implementation of Deep Network Accelerators Towards Healthcare and Biomedical Applications

  • king
  • (0) Download
  • 20210505
  • Save

... pages left unread,continue reading

Document pages: 23 pages

Abstract: The advent of dedicated Deep Learning (DL) accelerators and neuromorphicprocessors has brought on new opportunities for applying both Deep and SpikingNeural Network (SNN) algorithms to healthcare and biomedical applications atthe edge. This can facilitate the advancement of medical Internet of Things(IoT) systems and Point of Care (PoC) devices. In this paper, we provide atutorial describing how various technologies including emerging memristivedevices, Field Programmable Gate Arrays (FPGAs), and Complementary Metal OxideSemiconductor (CMOS) can be used to develop efficient DL accelerators to solvea wide variety of diagnostic, pattern recognition, and signal processingproblems in healthcare. Furthermore, we explore how spiking neuromorphicprocessors can complement their DL counterparts for processing biomedicalsignals. The tutorial is augmented with case studies of the vast literature onneural network and neuromorphic hardware as applied to the healthcare domain.We benchmark various hardware platforms by performing a sensor fusion signalprocessing task combining electromyography (EMG) signals with computer vision.Comparisons are made between dedicated neuromorphic processors and embedded AIaccelerators in terms of inference latency and energy. Finally, we provide ouranalysis of the field and share a perspective on the advantages, disadvantages,challenges, and opportunities that various accelerators and neuromorphicprocessors introduce to healthcare and biomedical domains.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×