Cervical Cancer Cell Detection Based on Deep Convolutional Neural Network
Mingyang Xia, Guoshan Zhang, Chaoxu Mu, Bin Guan, Mengxuan Wang
Abstract
The precise location of cancerous cells in thousands of cervical squamous epithelial cells can effectively reduce the workload of doctors and improve the accuracy of cervical cancer diagnosis. In this paper, we propose a new network structure for cervical cancer cells detection, named Series-parallel fusion network (SPFNet). Compared with traditional frameworks that use classification networks as the backbone for image feature extraction, we use different combination strategies in the series module and design five different head components to find the most suitable network structure for the detection task. In addition, data preprocessing such as RoI sliding window clipping is carried out for the thinprep cytologic images in cervical cancer. In order to compare the proposed framework with the state-of-the-art detection, we test these object detection algorithms on the same cervical cancer dataset. The experimental results show that our detection framework generates the optimum performances better than any others.