RESUMO
The sequential changes in shape of left ventricle (LV), which are the result of cellular interactions and their levels of organizational complexity, in its long axis view during one cardiac cycle are obtained. The changes are presented in terms of shape descriptors by processing of images obtained from a normal subject and two patients with dilated left ventricular cardio-myopathy. These images are processed, frame by frame, by a semi-automatic algorithm developed by MATLAB. This is consisting of gray scale conversion, the LV contour extraction by application of median and SRAD filters, and morphological operations. By filling the identified region with pixels and number of pixels along its contour the area and perimeter are calculated, respectively. From these the changes in LV volume and shape index are calculated. Based on these the stroke volume (SV) and ejection fraction (EF) are calculated. The changes in LV area, perimeter, volume and shape index in cardiac patients are less than that of normal subject. The calculated SV and EF of normal subject are within the range as obtained by various imaging procedures.