Detection of inferior myocardial infarction using shallow convolutional neural networks

Reasat, Tahsin; Shahnaz, Celia

doi:10.1109/r10-htc.2017.8289058

Cited by 57 publications

(54 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, in the case of Myocardial Infarction detection, the proposed approach achieves an accuracy of 86% with sensitivity and specificity of 96% and 84% respectively. Note that, the accuracy and specificity scores are similar to the state-of-the-art reported in [9], while the sensitivity is superior by 11%. We observe similar results in other channel configuration and disease predictions.…”

Section: B Resultssupporting

confidence: 70%

“…Consequently, neural network based solutions have been developed for this problem, in particular for detecting Myocardial Infarction. Similar to [8], [9], in this paper, we consider an additional challenge that it is necessary to perform predictions using a limited channel configuration at test time. We show that, RNNs are plagued by overfitting, and they produce inferior predictions under the limited channel setting.…”

Section: A Problem Statementmentioning

confidence: 99%

“…Denoting the multi-variate sequence dataset as X ∈ R N ×T ×K , where N denotes the number of training samples, T denotes the number of time-steps in each measurements and K indicates the total number of channels. The channel configuration is typically determined based on the disease to be diagnosed, for example, the channels {II, III, aVF} are known to be essential for detecting Myocardial Infarction [9]. Denoting the set of limited channels by C, whose cardinalityK K, we extract the matrixX ∈ R N ×T ×K .…”

Section: B Stage 1: Generative Models For Limited Channel Datamentioning

confidence: 99%

See 2 more Smart Citations

A Generative Modeling Approach to Limited Channel ECG Classification

Rajan

Thiagarajan

2018

2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

View full text Add to dashboard Cite

Processing temporal sequences is central to a variety of applications in health care, and in particular multichannel Electrocardiogram (ECG) is a highly prevalent diagnostic modality that relies on robust sequence modeling. While Recurrent Neural Networks (RNNs) have led to significant advances in automated diagnosis with time-series data, they perform poorly when models are trained using a limited set of channels. A crucial limitation of existing solutions is that they rely solely on discriminative models, which tend to generalize poorly in such scenarios. In order to combat this limitation, we develop a generative modeling approach to limited channel ECG classification. This approach first uses a Seq2Seq model to implicitly generate the missing channel information, and then uses the latent representation to perform the actual supervisory task. This decoupling enables the use of unsupervised data and also provides highly robust metric spaces for subsequent discriminative learning. Our experiments with the Physionet dataset clearly evidence the effectiveness of our approach over standard RNNs in disease prediction.

show abstract

Section: B Resultssupporting

confidence: 70%

Section: A Problem Statementmentioning

confidence: 99%

Section: B Stage 1: Generative Models For Limited Channel Datamentioning

confidence: 99%

See 1 more Smart Citation

A Generative Modeling Approach to Limited Channel ECG Classification

Rajan

Thiagarajan

2018

2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

View full text Add to dashboard Cite

show abstract

“…The main contributions presented in this paper are the following: 1) We put forward a fully convolutional neural network for myocardial infarction detection on the PTB dataset [11], [12] focusing on the clinically most relevant case of 12 leads. It outperforms state-of-the-art literature approaches [13], [14] and reaches the performance level of human cardiologists reported in an earlier comparative study [15]. 2) We study in detail the classification performance on subdiagnoses and investigate channel selection and its clinical implications.…”

Section: Introductionmentioning

confidence: 56%

Detecting and interpreting myocardial infarction using fully convolutional neural networks

Strodthoff

Strodthoff²

2019

Physiol. Meas.

155

View full text Add to dashboard Cite

Objective: We aim to provide an algorithm for the detection of myocardial infarction that operates directly on ECG data without any preprocessing and to investigate its decision criteria. Approach: We train an ensemble of fully convolutional neural networks on the PTB ECG dataset and apply state-of-the-art attribution methods. Main results: Our classifier reaches 93.3% sensitivity and 89.7% specificity evaluated using 10-fold cross-validation with sampling based on patients. The presented method outperforms state-ofthe-art approaches and reaches the performance level of human cardiologists for detection of myocardial infarction. We are able to discriminate channel-specific regions that contribute most significantly to the neural network's decision. Interestingly, the network's decision is influenced by signs also recognized by human cardiologists as indicative of myocardial infarction. Significance: Our results demonstrate the high prospects of algorithmic ECG analysis for future clinical applications considering both its quantitative performance as well as the possibility of assessing decision criteria on a per-example basis, which enhances the comprehensibility of the approach.

show abstract

“…In our study, a total of 504 records from 250 patients were used. The ECG records were pre-processed to remove noise, reduce sampling rate, and divided into frames of about 3 seconds long prior to being parsed through the neural network, as described in [15]. The resulting dataset contains 18, 040 ECG frames downsampled to 64 Hz .…”

Section: Experiments Setupmentioning

confidence: 99%

Improving Generalization of Deep Models for Cardiac Disease Detection Using Limited Channel ECG

Rajan¹,

Beymer

Narayan

2018

Computing in Cardiology Conference (CinC)

View full text Add to dashboard Cite

Acceleration of machine learning research in healthcare is challenged by lack of large annotated and balanced datasets. Furthermore, dealing with measurement inaccuracies and exploiting unsupervised data are considered to be central to improving existing solutions. In particular, a primary objective in predictive modeling is to generalize well to both unseen variations within the observed classes, and unseen classes. In this work, we consider such a challenging problem in machine learning driven diagnosis detecting a gamut of cardiovascular conditions (e.g. infarction, dysrhythmia etc.) from limited channel ECG measurements. Though deep neural networks have achieved unprecedented success in predictive modeling, they rely solely on discriminative models that can generalize poorly to unseen classes. We argue that unsupervised learning can be utilized to construct effective latent spaces that facilitate better generalization. This work extensively compares the generalization of our proposed approach against a state-of-the-art deep learning solution. Our results show significant improvements in F1-scores.

show abstract

Detection of inferior myocardial infarction using shallow convolutional neural networks

Cited by 57 publications

References 27 publications

A Generative Modeling Approach to Limited Channel ECG Classification

A Generative Modeling Approach to Limited Channel ECG Classification

Detecting and interpreting myocardial infarction using fully convolutional neural networks

Improving Generalization of Deep Models for Cardiac Disease Detection Using Limited Channel ECG

Contact Info

Product

Resources

About