Identifying and Classifying an Ovarian Cyst using SCBOD (Size and Count-Based Ovarian Detection) Algorithm in Ultrasound Image

Abstract

Polycystic ovaries are a sign of increasing infertility in the female population worldwide. An excessive number of follicle formations leads to polycystic ovarian syndromes. It affects the female reproductive cycle and leads to disorders such as cardiovascular issues, diabetes mellitus, and cancer. Calculating the number of follicles and detecting the follicle size is still challenging due to time complexity. Since the size of follicles and the greater number mislead the detection of the ovarian type in the ultrasound image. The ultrasound images contain more speckle noise, making the ovarian follicles difficult to see manually. A new convenient method is proposed for the detection of follicles and ovary classification is based on the measurement of size and the count of each follicle. In this paper, the work is divided into four steps, the first step preprocessing the ultrasound image. In the second step, the segmentation process is applied for object selection and separation process using an improved watershed algorithm. In the third step, based on the geometrical and statistical features the object is recognized by SCBOD accurately using various parameters such as size, count, mean, standard deviation, etc., Finally, an SVM classifier is used for classification to conclude the Polycystic ovary syndrome(PCOS) and Non-PCOS. This algorithm is proposed to the physician to find the ovarian follicles rapidly, which will offer accurate performance and is more effective in execution by adopting the SCBOD (Size and Count-based Object Detection) method.