Performance evaluation of neutrosophic similarity score driven segmentation of fetal ventricular chambers using b mode ultrasound images
Year: 2026
Type: article
Abstract: Fetal echocardiography is generally used for screening during pregnancy to recognize the absence/presence of heart chambers, structural abnormalities. The anatomical defects of the heart chambers can be examined, provided the chambers are segmented appropriately. This proposed study presents the application of neutrosophic similarity score (NSS) driven segmentation along with level set for segmenting the ventricular chambers of the fetal heart images obtained from the ultrasonic cineloop sequences. Clinically significant diastole frames as master frames were automatically selected from the cineloop sequences using the frame similarity measure and then the segmentation technique was applied. NSS provides a better classification of pixels into different regions by introducing an additional state referred to as Intermediate state in addition to True and False states. This is possible since NSS computes the membership functions under different conditions such as using the intensity image, mean of the image and homogeneity value obtained by considering the output of the texture energy measurement filter. Due to the existence of intermediate states, NSS provides well delineated boundaries resulting in improved segmentation. The performance of the segmentation is assessed using area error and distance error metrics with that of the ground truth validation performed by the clinical experts and the results were compared with well-known segmentation techniques, level set and Markov Random Field. The proposed NSS-level set segmentation outperforms others techniques in terms of recognizing right and left ventricular chambers showing a sensitivity of 91.32%, specificity of 93.6%, accuracy of 92.6%, Dice Coefficient of 0.828 ± 0.064
Source: Results in Engineering
Authors: N Sriraam, TV Sushma, S Suresh
Institution: Dayananda Sagar College of Engineering
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,940
OpenAlex ID: W7135376808