Neural Networks Implementations on FPGA for Biomedical Applications: A Review

Neethu Mohan, Asmaa Hosni, Mohamed Atef

Research output: Contribution to journalReview articlepeer-review

Abstract

The use of artificial intelligence in healthcare applications offers significant accuracy and utility for medical practitioners and patients. Deep learning has made a substantial positive impact on the healthcare industry by reducing the use of medical resources and workloads. Neural networks (NNs), inspired by human brain neurons, have attracted growing demand in healthcare systems. While NNs provide good performance, they must also meet high computation requirements and other features such as low power consumption, high accuracy, flexibility, and reconfigurability. Additionally, cost, execution speed, and degradation factors are important. This work discusses and compares emerging technologies and hardware for NN implementations, focusing on the requirements for wearable biomedical devices. To address these challenges, field-programmable gate array (FPGA) implementation is adopted as a preferred solution. This work also provides a performance analysis and required resource utilization for different biomedical applications using NNs implemented on FPGAs.

Original languageEnglish
Article number1004
JournalSN Computer Science
Volume5
Issue number8
DOIs
Publication statusPublished - Dec 2024

Keywords

  • Biomedical signals
  • FPGA
  • Machine learning
  • Neural networks

ASJC Scopus subject areas

  • General Computer Science
  • Computer Science Applications
  • Computer Networks and Communications
  • Computer Graphics and Computer-Aided Design
  • Computational Theory and Mathematics
  • Artificial Intelligence

Fingerprint

Dive into the research topics of 'Neural Networks Implementations on FPGA for Biomedical Applications: A Review'. Together they form a unique fingerprint.

Cite this