This article describes the results of the evaluation of informative value of image features and application of machine learning methods to solving the problem of delineation of the left ventricle of the heart. It is shown that the problem of left ventricle delineation is reduced to the problem of binary classification of pixels on every frame of echocardiography movies. The expediency of the applying of the classification decision tree and naive Bayesian is justified. Assessments of the error proportion in the classification by decision tree and naive Bayes classifier are obtained (error proportion values were 5.72% and 4.8%, respectively).