Computational Framework of Inverted Fuzzy C-Means and Quantum Convolutional Neural Network Towards Accurate Detection of Ovarian Tumors

Abstract

Due to the advancements in the lifestyle, stress builds enormously among individuals. A few recent studies have indicated that stress is a major contributor for infertility and subsequent ovarian cancer among women of reproductive age. In view of this, the present study proposes a two-stage computational methodology to identify and segment the ovarian tumour and classify it as benign or malignant. Using computerized tomography images, the first stage involves image segmentation using inverted fuzzy c-Means clustering, and second stage consists of deep quantum convolutional neural network in order to detect the tumours. The efficacy of the proposed method is demonstrated using in-house clinically collected dataset by comparing the results with the state-of-the-art methods. The experimental results confirm that the proposed approach outperforms the existing fuzzy C means algorithm by achieving the average Jaccard score of (0.65, 0.84, 0.79) (min, max, avg) and Dice score of (0.70, 0.83, 0.77) (min, max, avg), classification result of 78% for benign and 70.03% for malignant tumours. The classification results using the variant of convolutional neural network (CNN) model ResNet16 are compared with the quantum convolutional neural networks (QCNN) and obtained the classification performance of 87.02% for benign and 79.4% for malignant tumours and 84.4% for benign and 77.03% for malignant tumours respectively.