Detection of Ovarian Follicle for Polycystic Ovary Syndrome in Ultrasound Images of Ovaries
Detection of Ovarian Follicle for Polycystic Ovary Syndrome in Ultrasound Images of Ovaries
Author : BEDY PURNAMA; ADIWIJAYA; UNTARI NOVIA WISESTY; fhira nitha Published on : ICoICT 2015
Abstract
In the female in their reproductive cycle, Polycystic Ovary Syndrome (PCOS) is endocrine disorders that occurred on the adult female. In this paper, an application to detect ovarian follicles for PCO using USG images of ovaries is designed. The first stage is preprocessing, which employs low pass filter, equalization histogram, binarization, and morphological processes, to produce binary follicle images. The next stage is segmentation including edge detection, labeling, and cropping of follicle images. In the subsequent stage, which is feature extraction using Gabor wavelet, the cropped follicle images are categorized into two groups of texture features: (1) Mean, (2) Mean, Entropy, Kurtosis, Skewness, and Variance. This results in 2 datasets prepared for classification process, i.e. (1) dataset A, 40 images containing 26 normal images and 14 PCOS-indicated images, with Mean texture feature, obtained 275 follicle images, (2) dataset B, 40 images containing 34 normal images and 6 PCOS-indicated images, with Mean, Entropy, Kurtosis, Skewness, and Variance texture features, obtained 339 follicle images. The following stage is classification. It identifies the features of PCO and non-PCO follicles based on the feature vectors resulted from feature extraction. Here, three classification scenarios are designed: (1) Neural Network-Learning Vector Quantization (LVQ) method, (2) KNN – euclidean distance, and (3) Support Vector Machine (SVM). The best accuracy gained from SVM classification with RBF Kernel and C=40, shows that dataset A reach 82.55% while dataset B by 78.81% is obtained from KNN-euclidean distance classification on K=5.
