Extraction of endocardial motion is important to provide more accurate myocardial motion for further cardiac image analysis. This process involves endocardial detection which is challenging due to the presence of speckle noise and discontinuities of endocardial boundary. Researchers actively investigate various approaches to perform endocardial detection even though they involve extensive computations task and time consuming. In this paper, we implement a basic edge map detection using gradient and cavity-center-based method to detect the endocardial boundary. This method produces a boundary which still has discontinuities and outlier coordinates. To enhance the boundary, we propose a method based on a B-Spline and statistical approach to remove the discontinous points along the endocardium on parasternal short axis view of cardiac. The proposed method produces results which give visual evidences that it is able to significantly enhance the endocardial boundary