Design and Analysis of High Performance FPGA Based Convolutional Neural Network Accelerator for Abnormal Heart Beat Detection

R. Ravichandran, S. Jayachitra, M. Udhayakumar, S. Manoj Kumar, N. M. Yasod, S. Kavin

2025

Abstract

Aim: This project develops a CNN accelerator, which is less power consuming, on an FPGA basis to detect the abnormal heartbeats from the ECG reading in real-time. This accelerator can be utilized in wearable medical devices. Methods and Materials: This study involves two groups. We have used two tools are Anaconda and Xilinx vivado software. Group 1 refers to a novel FPGA-based CNN accelerator for abnormal heartbeat detection with 12 samples, and Group 2 refers to a conventional method for heartbeat detection (using CPU-based processing) with 12 samples. The power is set at 1W - 5W with a 10ms per speed, and the accuracy value is 98.5%. Result: The FPGA-based CNN accelerator obtains 98.4% accuracy, with a 12.3 ms latency and 30% energy savings for real-time ECG analysis. At a speed of 1500 samples/sec, it utilizes parallelism and quantization for processing the samples. The future work should focus on improving the scalability and the hybridization of the FPGA-GPU integration. Conclusion: The FPGA-based CNN accelerator is ideal for wearable and remote cardiac monitoring because it offers real-time, low-latency, and energy-efficient abnormal heartbeat detection, outperforming CPU/GPU methods.

Paper Citation

in Harvard Style

Ravichandran R., Jayachitra S., Udhayakumar M., Kumar S., Yasod N. and Kavin S. (2025). Design and Analysis of High Performance FPGA Based Convolutional Neural Network Accelerator for Abnormal Heart Beat Detection. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 321-327. DOI: 10.5220/0013912300004919

in Bibtex Style

@conference{icrdicct`2525,
author={R. Ravichandran and S. Jayachitra and M. Udhayakumar and S. Kumar and N. Yasod and S. Kavin},
title={Design and Analysis of High Performance FPGA Based Convolutional Neural Network Accelerator for Abnormal Heart Beat Detection},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={321-327},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013912300004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25
TI - Design and Analysis of High Performance FPGA Based Convolutional Neural Network Accelerator for Abnormal Heart Beat Detection
SN - 978-989-758-777-1
AU - Ravichandran R.
AU - Jayachitra S.
AU - Udhayakumar M.
AU - Kumar S.
AU - Yasod N.
AU - Kavin S.
PY - 2025
SP - 321
EP - 327
DO - 10.5220/0013912300004919
PB - SciTePress