Cardiac Motion Evaluation for Disease Diagnosis Using ICA Basis Neural Network

摘要

In this paper, we used ICA based artificial neural network (ANN) and propose a robust technique for efficient 2D echocardiography image analyzing. This project can be divided in three parts. The first part is cardiac motion estimation from sequence of echocardiography. The second part is feature extraction from motion patterns using sources, these sources is derived from independent component analyzing (ICA) algorithm. The third part is the classification of normal and abnormal myocardial function using neural network.