An Efficient Water Strider Algorithm with Auto Encoder for Cervical Cancer Diagnosis using Pap Smear Images

C. Suguna,S. P. Balamurugan
DOI: https://doi.org/10.1109/icssit53264.2022.9716358
2022-01-20
Abstract:Cervical cancer is considered a major cause of death from cancer among women. The severity of cervical cancer can be avoided by the earlier identification of the diseases. Medical image processing and intelligent models acts a vital part in the investigation of cancerous cells. Due to the recent advancements of several technologies, computer aided diagnosis (CAD) models becomes inexpensive and consume less time. Owing to the tremendous success of deep learning (DL) models in several medical fields, this paper presents an efficient water strider algorithm with autoencoder for cervical cancer diagnosis (WSAAE-CCD) technique using pap smear images. The goal of the WS AAE-CCD technique is for detecting the existence of cervical cancer using pap smear images. The WSAAE-CCD technique contains adaptive filtering (AF) based pre-processing to eradicate the existence of noise present in it. Besides, kernelized fuzzy C-means (KFCM) based segmentation approach was derived for determining the affected cervical cells and Residual Network (ResNet5) model is applied as a feature extractor. Finally, AE based classification model is employed and the optimal parameter tuning of the AE model takes place using the WSA. For inspecting the improved cervical cancer classification performance of the WS AAE-CCD technique, a detailed experimental analysis is made and the outcomes pointed out the betterment of the WSAAE-CCD technique.
What problem does this paper attempt to address?