Improving the Prediction and Classification of PCOS using SCBOD Feature Extraction with Augmentation

To implement the proposed rule-based algorithm SCBOD(Size and Count based object detection algorithm with augmentation) in ultrasound images for finding the PCOS(Poly Cystic Ovary Syndromes) also known as Poly Cystic Ovary Disease in the ovary. It exists with the symptoms and signs of androgen excess and abnormal ovarian functions which leads to failure of the ovulation process. PCOS is a common hormonal change disorder that affects the endocrine system in the female reproductive system. It causes multi-genetic disorders including environmental influence, food habits, and other life-threatening issues. A new emerging trending technique is used to analyze the ultrasound images to recognize the different types of ovaries like Normal ovary, Cystic ovary, and PCOS. An improved novel SCBOD architecture is implemented to identify the ovary and classify the ovaries as polycystic ovaries or non-polycystic ovaries. In this paper, the work is divided into three methods, I. Ovary can detection and classification using CNN method with augmentation, II. Proposed SCBOD feature extraction and classification with SVM classifier, and III. Augmentation techniques with SCBOD feature extraction and Classification with SVM classifier. The proposed algorithm gives more accuracy when augmented the dataset and all the other methods by increasing the time complexity and performance, which are evaluated using geometrical, statistical, and other metrics. The pathologist can able to detect PCOS accurately with the help of the proposed novel SCBOD algorithm.