Semi-supervised Cell Recognition under Point Supervision

TL;DR

This paper introduces a semi-supervised cell recognition framework that enhances end-to-end models using pseudo labels, co-teaching, and distribution alignment, significantly improving accuracy and efficiency in histopathology image analysis.

Contribution

It pioneers the integration of semi-supervised learning with end-to-end cell recognition models, addressing accuracy and speed limitations of previous density map-based methods.

Findings

Improved recognition accuracy across four histopathology datasets.

Effective reduction of annotation costs through semi-supervised learning.

Robustness to different staining styles demonstrated.

Abstract

Cell recognition is a fundamental task in digital histopathology image analysis. Point-based cell recognition (PCR) methods normally require a vast number of annotations, which is extremely costly, time-consuming and labor-intensive. Semi-supervised learning (SSL) can provide a shortcut to make full use of cell information in gigapixel whole slide images without exhaustive labeling. However, research into semi-supervised point-based cell recognition (SSPCR) remains largely overlooked. Previous SSPCR works are all built on density map-based PCR models, which suffer from unsatisfactory accuracy, slow inference speed and high sensitivity to hyper-parameters. To address these issues, end-to-end PCR models are proposed recently. In this paper, we develop a SSPCR framework suitable for the end-to-end PCR models for the first time. Overall, we use the current models to generate pseudo labels for unlabeled images, which are in turn utilized to supervise the models training. Besides, we introduce a co-teaching strategy to overcome the confirmation bias problem that generally exists in self-training. A distribution alignment technique is also incorporated to produce high-quality, unbiased pseudo labels for unlabeled data. Experimental results on four histopathology datasets concerning different types of staining styles show the effectiveness and versatility of the proposed framework. Code is available at \textcolor{magenta}{\url{https://github.com/windygooo/SSPCR}