Detection and reconstruction of cavity contours in a single frame 2D echocardiography using periodic splines approximation on a roughly pre-processed image, and automatic tuning of the parameters of the model

The authors have applied Tikhonov theory of regularization to single frame 2D echocardiopediatric images, in order to reconstruct the cavity contour in real time. Once the problem is regularized, the authors approximate the contour using periodic splines, on a sample extracted by a radial seek algorithm for which the initialisation point is given by the operator. As the reconstruction of the contour is fast, the authors propose to use the cross-validation algorithm to tune automatically the parameters of this model. Finally the authors compare their results to the drawings of several physicians