A systematic review and Meta-data analysis on the applications of Deep Learning in Electrocardiogram

Abstract

The success of deep learning over the traditional machine learning techniques in handling artificial intelligence application tasks such as image processing, computer vision, object detection, speech recognition, medical imaging and so on, has made deep learning the buzz word that dominates Artificial Intelligence applications. From the last decade, the applications of deep learning in physiological signals such as electrocardiogram (ECG) have attracted a good number of research. However, previous surveys have not been able to provide a systematic comprehensive review including biometric ECG based systems of the applications of deep learning in ECG with respect to domain of applications. To address this gap, we conducted a systematic literature review on the applications of deep learning in ECG including biometric ECG based systems. The study analyzed systematically, 150 primary studies with evidence of the application of deep learning in ECG. The study shows that the applications of deep learning in ECG have been applied in different domains. We presented a new taxonomy of the domains of application of the deep learning in ECG. The paper also presented discussions on biometric ECG based systems and meta-data analysis of the studies based on the domain, area, task, deep learning models, dataset sources and preprocessing methods. Challenges and potential research opportunities were highlighted to enable novel research. We believe that this study will be useful to both new researchers and expert researchers who are seeking to add knowledge to the already existing body of knowledge in ECG signal processing using deep learning algorithm.

Supplementary information: The online version contains supplementary material available at 10.1007/s12652-022-03868-z.

Keywords: Biometric Electrocardiogram System; Deep learning; Driving; Electrocardiogram; Machine learning.

Fig. 1

Google Trends for “electrocardiogram + ECG” compared with “deep learning + deep neural networks” from January, 2010 to May, 2020; (a) Occurrence timeline graph and (b) prevalence occurrence by country sorted by interest for “deep learning + deep neural networks”

Fig. 2

General Organization of the Complete Review

Fig. 3

A Venn diagram showing relationship between AI, ML, ANN and DL

Fig. 4

A Neuron vs. Perceptron

Fig. 5

Typical artificial neural network with backpropagation processes

Fig. 6

Shows comparison between ANN and DL. A typical ANN has three layers including input, hidden and output layers. A DL has at least input layer, > 2 hidden layers and output layer

Fig. 7

The typical DNN Structure

Fig. 8

The architecture of CNN (Sengupta et al. 2020)

Fig. 9

Illustrates convolution operation with a stride of 1, a kernel size of 3 × 3 and no padding (Yamashita et al. 2018)

Fig. 10

Architecture of LSTM (Tobore et al. 2019)

Fig. 11

Deep Belief Network Structure

Fig. 12

Architecture of Autoencoder (Sengupta et al. 2020)

Fig. 13

The Architecture of GAN

Fig. 14

The ECG signal components

Fig. 15

The structure of the human heart

Fig. 16

Literature retrieval, selection and evaluation for inclusion processes

Fig. 17

(a) Distribution of papers in databases across year category (2010–2020) (b) Showing type of publication

Fig. 18

The number of DL applications in ECG

Fig. 19

Domain and application area of the proposed DL models for ECG

Fig. 20

Taxonomy of the DL model application tasks

Fig. 21

Number of dataset sources used in the proposed DL models

Fig. 22

Showing number of deep neural networks’ role in training architecture