A Deep Learning Pipeline for Cervical Cancer Detection

Cervical cancer is a significant health issue, and early detection is crucial for increasing the chances of successful treatment. In recent years, research has focused on developing an automated system to detect cervical cancer at an early stage. This study proposes an automated system that uses feature extractors and machine learning classifiers to detect cervical cancer. The proposed system uses different types of feature extractors such as ResNet, DenseNet, and InceptionV3, and other extractors were tested to determine the optimal performance. Also, different filter techniques have been explored, including green filter extraction, blue filter extraction, etc., to enhance the feature extraction process. The proposed system achieved the highest accuracy of 86.48% using the DenseNet201 and DenseNet169 feature extractors in conjunction with a machine learning classifier known as Linear SVM. This accuracy level indicates that the proposed system has great potential as a tool to assist doctors and researchers in the early detection of cervical cancer. Overall, the proposed system provides an automated, accurate, and efficient method for detecting cervical cancer. This method has the potential to help doctors and researchers in diagnosing cervical cancer and ultimately improve patient outcomes. Further research could focus on improving the performance of the system, using larger datasets, and testing the system in real-world clinical settings.