Mid-diastole Frame Detection using Echocardiography Images
Applied image processing and ML algorithms to detect mid-diastole frames in echocardiography for mitral valve stenosis diagnosis.
Mid-diastole Frame Detection using Echocardiography Images
This project aimed to assist in the diagnosis of mitral valve stenosis by detecting the mid-diastole frame in echocardiography image sequences. By combining classical image processing with machine learning techniques, the system automatically identifies the optimal frame for analysis.
Description:
Applied image processing and machine learning algorithms to detect the mid-diastole frame in echocardiography images. The system supports diagnostic workflows by highlighting the most clinically relevant frame for measuring mitral valve area.
Tools & Technologies:
MATLAB, K-Means Clustering, Circular Hough Transform
Outcome:
Achieved an average frame difference of 1.40 frames compared to expert-selected frames (gold standard by echocardiologists), showing strong agreement and clinical viability.
System Overview
