Cervical Cancer Detection via Hybrid Particle Swarm Optimization and Hybrid Vision Transformer: A Dual-Innovation Deep Learning Method for Pap Smear Images
Abstract
Cervical cancer is a significant problem that affects the health of people all over the world and is responsible for a significant number of fatalities attributable to cancer in females. Because early detection is so important in terms of reducing death rates, screening processes need to be as accurate and efficient as possible. In recent years, there has been a growth in the popularity of deep learning approaches due to their capability to learn complex patterns and features from enormous datasets. This research study introduces a novel dual-innovation deep learning framework for malignant cell prediction using Pap smear test images. The framework integrates Hybrid Particle Swarm Optimization (HPSO) for advanced image enhancement and a Hybrid Vision Transformer (HViT) architecture for robust classification. The HViT model uniquely combines the local feature extraction capabilities of Convolutional Neural Networks (CNNs) with the global context modeling strengths of Vision Transformers. The proposed methodology effectively segregates cell regions from input images and demonstrates superior performance, achieving 98.1% accuracy across all classes over the maximum number of epochs. Furthermore, the system provides precise classification of the testing image, offering a significant advancement in automated cervical cancer diagnosis.