Automated analysis of gestational sac in medical image processing

Abstract

Ultrasonography is considered to be one of the most powerful techniques for imaging organs for an obstetrician and gynecologist. The first trimester of pregnancy is the most critical period in human existence. This evaluation of the first trimester pregnancy is usually indicated to confirm presence and number of pregnancy, its location and confirm well being of the pregnancy. The first element to be measurable is the gestational sac(gsac) of the early pregnancy. Size of gestational sac gives measure of fetus age in early pregnancy and also from that EDD is predicted. Today, the monitoring of gestational sac is done non-automatic, with human interaction. These methods involve multiple subjective decisions which increase the possibility of interobserver error. Because of the tedious and time-consuming nature of manual measurement, an automated, computer-based method is desirable which gives accurate boundary detection, consequently finding accurate diameter. Ultrasound images are characterized by speckle noise and edge information, which is weak and discontinuous. Therefore, traditional edge detection techniques are susceptible to spurious responses when applied to ultrasound imagery due to speckle noise. Algorithm for finding edges of gsac are as follows. In first step, we are using contrast enhancement, followed by filtering. We are smoothing image using lowpass filter followed by wiener filter. This image is segmented using thresholding. This results in image having large number of gaps due to high intensity around sac. These false regions are minimized by morphological reconstruction. Then boundaries are detected using morphological operations. Knowledge based filtering is used to remove false boundaries. In this prior knowledge of shape of gestational sac is used. First fragmented edges are removed then most circular shape is found as our sac is generally circular. Once sac is located, sac size is measured to predict the gestational age.

Computer-Aided Monitoring of Fetus Health From Ultrasound Images: A Review

“…Traditional Image Processing methods (Chakkarwar, Joshi & Revankar, 2010;Wee et al, 2010;Rawat et al, 2013;Ibrahim et al,2017;Lu et al, 2005;Satwika et al, 2013;Foi et al, 2014;Sahili et al, 2019;Rasheed et al, 2021;Thomas, Peters & Jeanty, 1991;Mukherjee et al, 2010;Wang, 2014;Methews & Deepa, 2014;Khan et al, 2015;Yu, Wang & Chen, 2015;Amoah, Anto & Crimi, 2015;Ravishankar, Prabhu & Vaidya, 2016;Hermawati et al, 2019;Yu et al, 2008;Wang et al, 2014;Jang et al, 2018;Kim et al, 2018;Nirmala & Palanisami, 2010;Lakshmi et al, 2008;Karl, Kagan & Chaoui, 2012;Sonia & Shanti, 2015;Deng et al, 2012;Anzalone et al, 2013;Sonia & Shanti, 2016;Nie et al, 2017) Supervised Learning methods (Khazendar et al, 2014;Ibrahim et al, 2017;Lu et al, 2005;Sahli et al, 2019;Carneiro et al, 2008;Anjit & Rishidas, 2011;Rafeek & Gunasundari, 2013;Deng et al, 2012;Park et al, 2013;Sciortino, Tegolo & Valenti, 2017) Unsupervised Learning methods…”

“…The obtained femur bone was used to create a single pixel-wide skeleton and finally FL measurement was made. Chakkarwar et al (2010) worked on automatic measurement of GS size. In the study, segmentation was made with thresholding technique using Gaussian and Wiener filters for speckle noise, respectively, and the diameter of the GS was measured.…”

Automated Techniques for the Interpretation of Fetal Abnormalities: A Review

“…The gestational sac diameter has been used as the first fetal parameter for confirmation of pregnancy. Chakkarwar et al [ 15 ] worked for finding the diameter of G.Sac. In this work, two steps are followed: in the first step, the global thresholding technique was used [ 16 ], then in the second step the diameter of G.Sac was measured.…”