A contour extraction algorithm for echocardiographic images

Identification of the cardiac borders is a basic goal in the analysis of echocardiographic images. This is usually accomplished by first applying some form of edge detection algorithm to identify a set of generic edge points, then extracting the target border from the edge map. A commonly used method of border identification is minimization of a cost function. In practical terms, the minimization problem can be solved only by limiting the domain of possible border configurations. This is usually accomplished by manually specifying an initial border approximation or region of interest. Here, the authors describe a cost function based border extraction method which does not require an initial border approximation. Minimal operator input is required; in most cases only two points must be specified. The algorithm then extracts the border by means of an efficient search procedure.