Detection of Cervical Cancer Cells in Whole Slide Images Using Deformable and Global Context Aware Faster RCNN-FPN

Abstract

Cervical cancer is a worldwide public health problem with a high rate of illness and mortality among women. In this study, we proposed a novel framework based on Faster RCNN-FPN architecture for the detection of abnormal cervical cells in cytology images from a cancer screening test. We extended the Faster RCNN-FPN model by infusing deformable convolution layers into the feature pyramid network (FPN) to improve scalability. Furthermore, we introduced a global contextual aware module alongside the Region Proposal Network (RPN) to enhance the spatial correlation between the background and the foreground. Extensive experimentations with the proposed deformable and global context aware (DGCA) RCNN were carried out using the cervical image dataset of “Digital Human Body” Vision Challenge from the Alibaba Cloud TianChi Company. Performance evaluation based on the mean average precision (mAP) and receiver operating characteristic (ROC) curve has demonstrated considerable advantages of the proposed framework. Particularly, when combined with tagging of the negative image samples using traditional computer-vision techniques, 6–9% increase in mAP has been achieved. The proposed DGCA-RCNN model has potential to become a clinically useful AI tool for automated detection of cervical cancer cells in whole slide images of Pap smear.

Automated Detection and Classification of Cervical Cancer Using Pap Smear Microscopic Images: A Comprehensive Review and Future Perspectives

“…6, 7 and 8 display the sample images of some of the publicly available and individually collected KMC dataset. [16,17]…”

“…Medical Image analysis includes different techniques and processing methods that aids abundantly in clinical analysis and interventions of medical images. Recently, in medical domains large volume of data and images can be analyzed and [13] 2003 500 Single Cell Images DTU/Herlev Dataset [13] 2005 917 Single Cell Images ISBI14 [14,15] 2014 961 (16 EDF + 945 Synthetic) Slide Images ISBI15 [14,15] 2015 37 (17 EDF (each with 20 FOVs) Slide Images CERVIX [14,15] 2018 113 (93 EDF (each with 20 FOVs) Slide Images CRIC Cervix [16] 2021 400 with 11,534 classified cells, 0.228 µm/pixel Slide Images SIPakMed [16] 2018 966 with 4049 classified cells Slide Images DHB [16] 2021 500 with 5414 classified cells Whole Slide Image interpreted by using different machine learning algorithms. These algorithms include statistical, probabilistic and different optimization techniques and allow computer to learn and analyze samples in a better way.…”

“…[73] Deep learning method outperforms in recognizing patterns and high-level sematic features in large volumes of data, extracting relationships between complex features in the images, and identifying characteristics that is difficult be perceived by the human brain. [16,73] [16,73]…”

“…X Li et al propose a novel framework based on "Faster RCNN-FPN architecture" [16] for the detection of abnormal cervical cells. This model can be used as AI tool for automated detection of cervical cancer cells in whole slide images of pap smear.…”

“…This model can be used as AI tool for automated detection of cervical cancer cells in whole slide images of pap smear. [16] Here authors introduces deformable convolution layer into feature extraction pyramid network resulting in the accurate detection of deformed cells of various sizes. Spatial context information also plays a very important role in the detection of abnormal cells.…”