Cervical Cytology Classification Using PCA and GWO Enhanced Deep Features Selection

Cervical cancer is one of the most deadly and common diseases among women worldwide. It completely curable if diagnosed in an early stage, but tedious costly detection procedure makes it unviable to conduct population-wise screening. Thus, augment effort clinicians, this paper, we propose a fully automated framework that utilizes deep learning feature selection using evolutionary optimization for cytology image classification. The proposed extracts from several convolution neural network (CNN) models uses two-step reduction approach ensure computation cost faster convergence. features extracted CNN form large space whose dimensionality reduced principal component analysis while preserving 99% variance. A non-redundant, optimal subset selected algorithm, grey wolf optimizer, thus improving classification performance. Finally, used train support vector machine classifier generating final predictions. evaluated on three publicly available benchmark datasets: Mendeley Liquid Based Cytology (4-class) dataset, Herlev Pap Smear (7-class) SIPaKMeD (5-class) dataset achieving accuracies 99.47, 98.32 97.87%, respectively, justifying reliability approach. relevant codes can be found in: https://github.com/DVLP-CMATERJU/Two-Step-Feature-Enhancement .