Arrhythmia Detection

This repository contains the implementation of a neural network for detecting arrhythmia using the dataset from Kaggle.

1. Model Architecture

The model architecture is designed to efficiently classify heartbeat signals into different classes, leveraging a combination of convolutional and residual blocks.

2. Residual Inverted Convolution (RIC) Block

The Residual Inverted Convolution (RIC) Block is a key component of the network, enhancing feature extraction through residual connections and inverted convolutions.

3. Results

The model has a total of 19.6K parameters. The performance metrics for each class are as follows:

Class	Precision	Recall	Specificity	NPV	Accuracy
Class N	0.9887	0.9966	0.9452	0.9829	0.9877
Class S	0.9161	0.7464	0.9982	0.9934	0.9918
Class V	0.9666	0.9606	0.9977	0.9972	0.9952
Class F	0.8897	0.7469	0.9993	0.9981	0.9974
Class Q	0.9925	0.9882	0.9994	0.9991	0.9986

Definitions:

• Precision: The ratio of true positive predictions to the total predicted positives.
• Recall: The ratio of true positive predictions to the total actual positives.
• Specificity: The ratio of true negative predictions to the total actual negatives.
• NPV (Negative Predictive Value): The ratio of true negative predictions to the total predicted negatives.
• Accuracy: The ratio of correct predictions (both true positives and true negatives) to the total predictions.

Dataset

The dataset used in this project can be found on Kaggle: Heartbeat Dataset.

4. References

Li, Duo, et al. "Involution: Inverting the inherence of convolution for visual recognition." Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2021.