Semi-supervised Semantic Segmentation for Effusion Cytology Images

Abstract

Cytopathologists analyse images captured at different magnifications to detect the malignancies in effusions. They identify the malignant cell clusters from the lower magnification, and the identified area is zoomed in to study cell level details in high magnification. The automatic segmentation of low magnification images saves scanning time and storage requirements. This work predicts the malignancy in the effusion cytology images at low magnification levels such as $$10{\times }$$ and 4 $$\times $$ . However, the biggest challenge is the difficulty in annotating the low magnification images, especially the 4 $$\times $$ data. We extend a semi-supervised learning (SSL) semantic model to train unlabelled 4 $$\times $$ data with the labelled 10 $$\times $$ data. The benign F-score on the predictions of 4 $$\times $$ data using the SSL model is improved 15% compared with the predictions of 4 $$\times $$ data on the semantic 10 $$\times $$ model.