Cervical Glandular Cell Detection from Whole Slide Image with Out-Of-Distribution Data

TL;DR

This paper introduces PolarNet, a novel model that leverages morphological prior knowledge and self-attention to reduce false positives in cervical glandular cell detection from whole slide images, improving accuracy and reliability.

Contribution

The paper proposes PolarNet, a plugin module utilizing morphological priors and self-attention to enhance glandular cell detection accuracy in the presence of out-of-distribution data.

Findings

PolarNet reduces false positives in GC detection.

Integrating PolarNet improves detection accuracy by up to 8.8%.

General models' mean average precision increases by 0.007 to 0.015.

Abstract

Cervical glandular cell (GC) detection is a key step in computer-aided diagnosis for cervical adenocarcinomas screening. It is challenging to accurately recognize GCs in cervical smears in which squamous cells are the major. Widely existing Out-Of-Distribution (OOD) data in the entire smear leads decreasing reliability of machine learning system for GC detection. Although, the State-Of-The-Art (SOTA) deep learning model can outperform pathologists in preselected regions of interest, the mass False Positive (FP) prediction with high probability is still unsolved when facing such gigapixel whole slide image. This paper proposed a novel PolarNet based on the morphological prior knowledge of GC trying to solve the FP problem via a self-attention mechanism in eight-neighbor. It estimates the polar orientation of nucleus of GC. As a plugin module, PolarNet can guide the deep feature and predicted confidence of general object detection models. In experiments, we discovered that general models based on four different frameworks can reject FP in small image set and increase the mean of average precision (mAP) by $\text{0.007}\sim\text{0.015}$ in average, where the highest exceeds the recent cervical cell detection model 0.037. By plugging PolarNet, the deployed C++ program improved by 8.8\% on accuracy of top-20 GC detection from external WSIs, while sacrificing 14.4 s of computational time. Code is available in https://github.com/Chrisa142857/PolarNet-GCdet