Automated delineation of Doppler echocardiographic images using texture filters

The manually traced continuous wave Doppler (CWD) echocardiographic images are used in estimating the severity of valvular abnormalities. Manual tracing is a time consuming activity embedded with inter- and intra-personal variations. To address this issue an automatic delineation technique based on texture filters is proposed in this work. The outer boundaries of the CWD images reflecting aortic regurgitation are delineated. The images are traced in two stages using the combination of range, standard deviation and entropy filters. The automatic delineation of CWD images will result in the faster processing accompanied with uniform and accurate clinical conclusions. The performance of proposed delineation techniques are compared with two state-of-the art segmentation techniques. Visual analyses of results reveal that the texture filters combination of the range filter and the standard deviation filter with entropy filter, respectively result in the best delineated outer boundary.