Self-attention based ResNet model for Cervical Cancer Detection

One of the life taking disease in the trait of cancer is Cervical cancer. It is very sensitive case not only to females but also to males. Therefore, due to lack of comfort to talk about in society and unavailability of medial equipments for treatment bring in the need for computation in action to help the society with correct prediction of cervical cancer cases based on the cervix images and detecting the major area of focus. This study advances a unique method for detecting cervical cancer using images from the ICAR-WHO dataset overcoming a number of difficulties, such as the lack of labeling and dataset availability. To address these challenges, the proposed methodology employs convolution neural networks along with clustering technique to automatically detect and classify cervical cancer images using pseudo labeling. The work relies on standard ResNet50 model and encourages self-learning of the features using attention mechanism for better detection and classification of cervical cancer images tested using different methods. Overall, the proposed method demonstrates promising results for cervix cancer detection and holds potential for improved diagnosis and treatment planning.