Learning with Less Labels in Digital Pathology via Scribble Supervision from Natural Images

TL;DR

This paper demonstrates that using scribble supervision from natural images can significantly improve deep learning models in digital pathology, reducing annotation effort while maintaining high performance.

Contribution

It introduces the use of scribble labels from natural images to enhance digital pathology models, showing comparable results to full segmentation labels with less annotation effort.

Findings

Scribble labels boost digital pathology model performance.

Models with scribble supervision match full segmentation label results.

Scribble annotation is faster and easier to collect.

Abstract

A critical challenge of training deep learning models in the Digital Pathology (DP) domain is the high annotation cost by medical experts. One way to tackle this issue is via transfer learning from the natural image domain (NI), where the annotation cost is considerably cheaper. Cross-domain transfer learning from NI to DP is shown to be successful via class labels. One potential weakness of relying on class labels is the lack of spatial information, which can be obtained from spatial labels such as full pixel-wise segmentation labels and scribble labels. We demonstrate that scribble labels from NI domain can boost the performance of DP models on two cancer classification datasets (Patch Camelyon Breast Cancer and Colorectal Cancer dataset). Furthermore, we show that models trained with scribble labels yield the same performance boost as full pixel-wise segmentation labels despite being significantly easier and faster to collect.