Congenital Heart Defects (CHD) are widely responsible for more than 10% of neonatal mortality in India. The objective of this work describes the sequential application of several techniques such as motion estimation, speckle suppression, image enhancement, image alignment and edge detection to delineate the edge structures of the fetal heart from clinical ultrasound images. The prominent image processing techniques used to localize the edge structures of fetal heart are Probabilistic Patch Based Weighted Maximum Likelihood Estimation (PPBMLE) based speckle noise suppression technique, Shock filter image enhancement technique and proposed modified edge detection with Binary connectedness based ANT algorithm. The main contribution of the proposed ANT algorithm for edge detection uses binary connectedness property among pixel intensities to efficiently update the heuristic information of ANT algorithm. The quantitative performance merit prove that the sequential application of the above mentioned image processing techniques was predominantly useful in delineating the ultrasound fetal heart structures and interpret pathophysiologic biomarkers from images.
Keywords
Ant Colony Optimization, Binary Connectedness, Congenital Heart Disease, Detection Morphology, Edge
User
Information