Cervical Cancer Classification using a Novel Hybrid Approach

Jheelam Mondal¹Rajdeep Chatterjee¹Mahendra Kumar Gourisaria^1*Ernesto Leon-Castro^2*Manoj Sahni³

• ¹KIIT Deemed to be University, Bhubaneswar, India
• ²Universidad Catolica de la Santisima Concepcion, concepcion, Chile
• ³Pandit Deendayal Energy University, Gandhinagar, India

Cervical cancer is among the most frequently diagnosed malignancies in women. It is the fourth most prevalent malignancy in women worldwide. Pap smear tests, a popular and effective medical procedure, enable the early detection and screening of cervical cancer. Expert physicians perform the smear analysis, which is a laborious, time consuming and prone to mistakes. In the developing and underdeveloped nations, the low output of cervical cancer detection procedures is caused by a lack of resources and highly qualified healthcare professionals. Researchers were inspired to use deep learning (DL), which has recently gained popularity, to diagnose cervical cancer. Many deep learning-based methods have been created as a result. In this study, we propose a novel technique by combining the concept of feature extraction by multi-head self-attention blocks, cross-stage partial network and feature fusion integration by spatial pyramid pooling fast layer components to identify healthy and cancerous cervical cells. The experimental study of our proposed model CASPNet (Contextual Attention and Spatial Pooling Network) has achieved an accuracy of 97.07% in the widely used benchmark SIPAKMED dataset.