Atrium segmentation using machine learning

Soni, Yash and Jangirala, Srinivas (2022) Atrium segmentation using machine learning. In: 7th IEEE International Conference on Recent Advances and Innovations in Engineering, ICRAIE 2022, 1-3 December 2022, Mangalore.

[thumbnail of Atrium Segmentation Using Machine Learning.pdf] Text
Atrium Segmentation Using Machine Learning.pdf - Published Version
Restricted to Registered users only

Download (1MB) | Request a copy

Abstract

The segmentation and 3D reconstruction of the human atria are critical for the precise diagnosis and treatment of atrial fibrillation, the most common cardiac arrhythmia, as the manual segmentation of the atria from medical images is a fallible process. Therefore, an automated process is highly desirable. This research aims to create a segmentation pipeline that includes a convolutional neural network (CNN) based on the U-Net architecture. A dataset consisting of 20 MRI scans of the heart with a corresponding ground truth mask is provided by the medical segmentation decathlon. In a 2-dimensional setting, this translates to 4542 MRI scans of the heart and labeled slices. Add more details on the output and the inferences drawn out of our work. The article proposes a machine learning algorithm has been created which can assist in the segmentation of the atrium using machine learning. It explains the process and techniques that have been used in the processing and training stages of the model. Finally, the model result and future directions have been discussed.

Item Type: Conference or Workshop Item (Paper)
Keywords: Left Atrium Segmentation | MRI Scans | Machine Learning | Data Science | Training | Testing | Validation
Subjects: Physical, Life and Health Sciences > Engineering and Technology
JGU School/Centre: Jindal Global Business School
Depositing User: Amees Mohammad
Date Deposited: 23 Mar 2023 04:54
Last Modified: 23 Mar 2023 04:54
Official URL: https://doi.org/10.1109/ICRAIE56454.2022.10054328
URI: https://pure.jgu.edu.in/id/eprint/5722

Downloads

Downloads per month over past year

Actions (login required)

View Item
View Item