Cervix Visionator ELM: A Novel Approach to Early Detection of Cervical Cancer
Ashfaque Khowaja, Zou BeiJi, Xiaoyan Kui · 2024
Cervical cancer is one of the leading causes of mortality among women, but early detection and treatment can prevent its progression. Although Pap smear images are a prevalent method of cervical cancer screening, manual diagnosis is laborious and prone to error. The Cervix Visionator ELM, a novel automated computerized approach for detecting cervical cancer in Pap smear images, is introduced in this research article. We integrate the self-attention mechanism with EfficientNet, a state-of-the-art Convolutional Neural Network (CNN) architecture, and Vision Transformer (ViT) models to get deep-learned features from Pap smear images. The extracted features are then classified using an Extreme Learning Machine (ELM)-based classifier. Our model was assessed using the SIPaKMeD open dataset. In cervical cells, the Cervix Visionator ELM demonstrated an accuracy of 98.89%, 99.42% precision, 97.87% recall, and 98.76% F-measure. The findings of this study provide evidence for the soundness and efficacy of our model, establishing its superiority over the majority of current models utilized for classifying cervical cells.