Localization of cardiac cavities using hierarchically search in 2D grayscale ultrasound images

Fetal echocardiography is an important tool for diagnosis. This paper presents a method to provide automatic localization of cardiac structures in fetal echocardiography images, where the early diagnostics of heart congenital diseases can improve results from medical treatment. The interest structures are the four cardiac cavities. The method use a hierarchical search based on a windowing technique with mold matching to find the cardiac cavities. This mold is calculated using a probability density function that receives as parameter the gray level of a representative image and also uses spatial features of the images. The hierarchical search is performed: (i) first, the image is covered to locate the heart; (ii) then, based on the heart position, the desired structure is found. The comparison of the mold and the candidate image is made using the Euclidean Distance. One of the main characteristics of the method is its rotation invariance.