Follicle Detection Model on Ovarian Ultrasound Image

every woman has two ovaries. Ovaries have several follicles, which consist of oocytes or eggs which are filled with granulosa cells. Some women can have a difference in the number of follicles in each ovary. There are cases of several follicles that are coincided, making it difficult to calculate the number of follicles. In this study, the separation of adjoining follicles and automatic follicle counting was carried out from the results of ovarian ultrasound image segmentation. The segmentation results obtained feature information in the form of follicular feature extraction as many as eight features. The techniques used in this work for feature selection was carried out using Principal Components Analysis (PCA) to reduce the feature. In this study, the PCA and Support Vector Machine (SVM) classifier produced higher accuracy than the classification without PCA. The experimental results also show that the proposed method produced higher classification accuracy than previous work, which yielded 90.39% accuracy, 90.27 % sensitivity, and 90.43 % specificity.