Event-driven cardiac arrhythmia classification using antidictionaries
- Authors
- Publication Date
- Oct 14, 2024
- Source
- HAL
- Keywords
- Language
- English
- License
- Unknown
- External links
Abstract
Cardiovascular diseases are the leading cause of death globally, represent-ing 31.4% of all deaths in the world in 2022. With more than half of all cardiovascu-lar disease-related deaths happening in settings outside of the hospital, there is a needfor constant at-home monitoring to prevent them. Such long-term monitoring solutionsnecessitate autonomous, comfortable, and low-power consumption devices. Since elec-trocardiogram signals are constituted by nature by sparse events in time, a non-uniformquantization scheme is well suited to code this signal. It will enable a data-dependentpower consumption concentrated during the activity periods.In this context, this thesis investigates features used to differentiatearrhythmic and healthy heartbeats used in cardiology. A particular focus is made onthe implementation of the feature extraction process in an event-driven system to bene-fit from the sparse nature of the signal. In addition, a study of classification methodsboth integrated and non-integrated is presented focusing on finding ways to implementartificial intelligence directly in the system.Combining an event-driven sampling scheme with on-chip artificial intelligence (AI) thusaims to achieve a system complexity that is lower by several orders of magnitude thanstate-of-the-art systems. Most event-driven systems utilize time-related features to clas-sify arrhythmia, but they often require a sampling clock, which counteracts the ben-efits of event-driven systems.In this thesis, we investigate a system where the time information contained in the sampled input data is extracted without a regular clock, using successive delays coarsely coding the time between event-driven samples.Using the output of the delays and the output of a Level-Crossing Analog-to-Digital Converter, an event-driven token is generated for each sample, holding coarse information on the shape of the electrocardiogram signal.Using those tokens as the input of a neural network, an average accuracy of 98% canbe achieved using the MIT-BIH database. However, because training a neural networkwith unbalanced datasets introduces a bias, a novel classification scheme based on antidictionaries is introduced. It consists of patterns created from negative informationabout healthy heartbeats, allowing the detection of arrhythmic heartbeats by trainingAI-based systems without introducing a bias. Antidictionary-based classification achieves a similar average accuracy of 98%, showing its ability to classify arrhythmia on par with state-of-the-art neural networks.The system implemented on FPGA necessitates 99% less memory to store the classifier, 1,024 logic comparisons instead of 512 multiplication operations, and 99% less accumulation com pared to the state of the art.This results in 25.1% less LUT and 34.5% less register used on the FPGA compared to the state of the art and does not require the implementation of any RAM. For those benefits, the overall classification accuracy was reduced by 1.03 percentage points compared to the state of the art with a classification accuracy of 97.97%.
