A new neural network algorithm to study myocardial reflected ultrasound for tissue characterization

It is pointed out that ultrasonic myocardial tissue characterization is a promising approach to diagnose myocardial infarction since the acoustic properties of myocardial tissue are altered by the disease. Echocardiographic images are digitized to study ultrasound backscatter for tissue characterization. Neural network analysis is a successful approach to tissue characterization. A structure consisting of a multilayer perceptron combined with the feedback algorithm ALOPEX is proposed.<>