A Hybrid Model Combining CNN-SVM for Cervical Cancer Classification
Shaik Jaffar Hussain, Madapuri Rudra Kumar, K. Sailaja, R Venkataramana, Pole Anjaiah, M. Sudhakara · 2024
Cervical cancer is the ranks as fourth most common type of cancer globally, and early diagnosis is essential for successful treatment. It ranks among the top causes of death among women and treats it swiftly with the best medical guidance available. Deep learning (DL) techniques have grown in popularity for image classification. Convolution Neural Network (CNN) is one of the most valuable methods for extracting features without employing handcrafted models, eventually leading to more helpful classification accuracy. This study aims to overcome and deal with the obstacles associated with cervical cancer detection by incorporating the Mendeley LBC dataset with a hybrid technique. Our proposed Hybrid method combines convolutional neural networks (CNN) and support vector machine (SVM) algorithms to classify cervical cancer effectively. Our proposed Hybrid model demonstrated remarkable performance, boasting an outstanding accuracy of 98.92%. Furthermore, it showed exceptional precision at 98.93%, recall of 98.91%, an outstanding F-score of 98.92%, and a commendable Matthews Correlation Coefficient (MCC) value of 0.987, suggesting a strong correlation between predicted and observed classifications. In our comparative analysis, our proposed model outperformed existing models based on their accuracy.