Particle swarm optimation based 2-dimensional randomized hough transform for fetal head biometry detection and approximation in ultrasound imaging

Abstract

One of the most profound use of ultrasound imaging is to generate the image of fetal during pregnancy. This paper will describe an ellipse detection approach to automatically detect and approximate the head size of the fetal. The method was developed using the Hough Transform techniques that have been modified and optimized by Particle Swarm Optimization (PSO). Experiments of the method are tested on synthetic and real ellipse image dataset. For real images, the detection was applied on 2D ultrasonography images to perform fetal head measurement to approximate the Head Circumference (HC) and Biparietal Diameter (BPD). Experiment result showed that the proposed method can perform ellipse detection in synthetic dataset with satisfactory result for noisy images with noise density up to 0.4 and able to perform the fetal head detection for real images with an averate hit rate of 0.654. This proposed method can also perform detection on images that have high degree of noise or incomplete ellipse images generated from the fetal objects.