Accurate Cardiac Structure Segmentation from Echocardiographic Images using a Deep Neural Network: A Comprehensive Evaluation

Author : Seetharaman AR

Abstract :Accurate segmentation of cardiac structures from echocardiographic images is crucial for reliable diagnosis and quantitative assessment of cardiac function. Manual segmentation is time-consuming and inconsistent. This paper presents a fully automated deep learning-based segmentation framework using the U-Net architecture, optimized for echocardiographic image analysis. The proposed approach is trained and validated on a locally collected dataset using a 70/20/10 split. It achieves a Dice Coefficient of 0.94 and Intersection over Union (IoU) of 0.91, demonstrating robust delineation of cardiac chambers. Comparative results show that the U-Net architecture outperforms traditional segmentation methods, confirming its potential for clinical integration

Keywords :Cardiac segmentation, Echocardiography, U-Net, Deep learning, Medical image analysis, Automation.

Conference Name :International Conference on Engineering & Technology (ICET-25)

Conference Place Vellore,India

Conference Date 16th Nov 2025