Search

Purpose: An Electrocardiogram (ECG) is the simplest and fastest bio-medical test that is used to detect any heart-related disease. ECG signals are generally stored in paper form, which makes it difficult to store and analyze the data. While capturing ECG leads from paper ECG records, a lot of background information is also captured, which results in incorrect data interpretation. Methods: We propose a deep learning-based model for individually extracting all 12 leads from 12-lead ECG images captured using a camera. To simplify the analysis of the ECG and the calculation of complex parameters, we also propose a method to convert the paper ECG format into a storable digital format. The You Only Look Once, Version 3 (YOLOv3) algorithm has been used to extract the leads present in the image. These leads are then passed on to another deep learning model which separates the ECG signal and background from the single-lead image. After that, vertical scanning is performed on the ECG signal to convert it into a 1-Dimensional (1D) digital form. To perform the task of digitalization, we used the pix-2-pix deep learning model and binarized the ECG signals. Results: Our proposed method was able to achieve an accuracy of 97.4 %. Conclusion: The information on the paper ECG fades away over time. Hence, the digitized ECG signals make it possible to store the records and access them anytime. This proves highly beneficial for heart patients who require frequent ECG reports. The stored data can also be useful for research purposes, as this data can be used to develop computer algorithms that are capable of analyzing the data.

Haemorrhaging of the brain is the leading cause of death in people between the ages of 15 and 24 and the third leading cause of death in people older than that. Computed tomography (CT) is an imaging modality used to diagnose neurological emergencies, including stroke and traumatic brain injury. Recent advances in Deep Learning and Image Processing have utilised different modalities like CT scans to help automate the detection and segmentation of brain haemorrhage occurrences. In this paper, we propose a novel implementation of an architecture consisting of traditional Convolutional Neural Networks(CNN) along with Graph Neural Networks(GNN) to produce a holistic model for the task of brain haemorrhage segmentation.GNNs work on the principle of neighbourhood aggregation thus providing a reliable estimate of global structures present in images. GNNs work with few layers thus in turn requiring fewer parameters to work with. We were able to achieve a dice coefficient score of around 0.81 with limited data with our implementation., Comment: 10 pages 6 figures 3 tables preprint