AI Article Synopsis

Article Abstract

Using smart wearable devices to monitor patients' electrocardiogram (ECG) for real-time detection of arrhythmias can significantly improve healthcare outcomes. Convolutional neural network (CNN) based deep learning has been used successfully to detect anomalous beats in ECG. However, the computational complexity of existing CNN models prohibits them from being implemented in low-powered edge devices. Usually, such models are complex with lots of model parameters which results in large number of computations, memory, and power usage in edge devices. Network pruning techniques can reduce model complexity at the expense of performance in CNN models. This paper presents a novel multistage pruning technique that reduces CNN model complexity with negligible loss in performance compared to existing pruning techniques. An existing CNN model for ECG classification is used as a baseline reference. At 60% sparsity, the proposed technique achieves 97.7% accuracy and an F1 score of 93.59% for ECG classification tasks. This is an improvement of 3.3% and 9% for accuracy and F1 Score respectively, compared to traditional pruning with fine-tuning approach. Compared to the baseline model, we also achieve a 60.4% decrease in run-time complexity.

Download full-text PDF

Source
http://dx.doi.org/10.1109/EMBC46164.2021.9630588DOI Listing

Publication Analysis

Top Keywords

edge devices
12
multistage pruning
8
cnn based
8
existing cnn
8
cnn models
8
pruning techniques
8
model complexity
8
cnn model
8
ecg classification
8
accuracy score
8

Similar Publications

Blood components play a crucial role in maintaining human health and accurately detecting them is essential for medical diagnostics. A cutting-edge sensor utilizing PCF revealed to precisely identify a wide range of blood components with WBCs (white blood cells), RBCs (red blood cells), HB (hemoglobin), platelets, and plasma. A numerical analysis was performed using COMSOL Multiphysics software to assess the capabilities of the sensor.

View Article and Find Full Text PDF

Edge contacts offer a significant advantage for enhancing the performance of semiconducting transition metal dichalcogenide (TMDC) devices by interfacing with the metallic contacts on the lateral side, which allows the encapsulation of all of the channel material. However, despite intense research, the fabrication of feasible electrical edge contacts to TMDCs to improve device performance remains a great challenge, as interfacial chemical characterization via conventional methods is lacking. A major bottleneck in explicitly understanding the chemical and electronic properties of the edge contact at the metal-two-dimensional (2D) semiconductor interface is the small cross section when characterizing nominally one-dimensional edge contacts.

View Article and Find Full Text PDF

Thrombosis, a major cause of morbidity and mortality worldwide, presents a complex challenge in cardiovascular medicine due to the intricacy of clotting mechanisms in living organisms. Traditional research approaches, including clinical studies and animal models, often yield conflicting results due to the inability to control variables in these complex systems, highlighting the need for more precise investigative tools. This review explores the evolution of thrombosis models, from conventional polydimethylsiloxane (PDMS)-based microfluidic devices to advanced hydrogel-based systems and cutting-edge 3D bioprinted vascular constructs.

View Article and Find Full Text PDF

Next generation bioelectronic medicine: making the case for non-invasive closed-loop autonomic neuromodulation.

Bioelectron Med

January 2025

SecondWave Systems Incorporated, Head Quarters, Minneapolis-Saint Paul, MN, 55104, USA.

The field of bioelectronic medicine has advanced rapidly from rudimentary electrical therapies to cutting-edge closed-loop systems that integrate real-time physiological monitoring with adaptive neuromodulation. Early innovations, such as cardiac pacemakers and deep brain stimulation, paved the way for these sophisticated technologies. This review traces the historical and technological progression of bioelectronic medicine, culminating in the emerging potential of closed-loop devices for multiple disorders of the brain and body.

View Article and Find Full Text PDF

Background: In medial mobile-bearing unicompartmental knee arthroplasty (MB-UKA), the position of the bearing does not correspond to the planned position which will increasing the risk of bearing dislocation. This study aimed to explore the relationship between the malposition of the femoral and tibial components and the phenomenon of bearing deviation using postoperative radiological measurements.

Methods: One hundred twenty patients who underwent mobile-bearing uni-compartmental knee arthroplasty (MB-UKA) at our hospital between January and August 2023 were enrolled in this retrospective study.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!