Database guided detection of anatomical landmark points in 3D images of the heart

Automated landmark detection may facilitate the examination and automatic analysis of three-dimensional (3D) echocardiograms. By detecting 3D anatomical landmark points, the standard anatomical views can be extracted automatically, for better standardized visualization. Furthermore, the landmarks can serve as an initialization for other analysis methods, such as segmentation. The described algorithm applies landmark detection in perpendicular planes of the 3D dataset. It exploits a database of expert annotated images, using an extensive set of Haar features for classification. The detection is performed using two cascades of Adaboost classifiers in a coarse to fine scheme. The method can detect landmarks accurately in the four-chamber (apex: 7.9±7.1mm, mitral valve center: 4.8±2.3mm) and two-chamber (apex: 7.1±6.7mm, mitral valve center: 5.2±2.8mm) views.