Fusing Domains for Tumor Detection in Frozen Section Histopathology Using Joint Feature Alignment and Self-Knowledge Distillation

초록

Robustness in neural networks is of paramount importance in the field of image classification. Domain shift remains a major challenge particularly prevalent in whole-slide image (WSI) data within the context of histopathology, which is exacerbated by differences in WSI scanners and the status of WSIs. This study focuses on the domain shift encountered in tumor detection in breast lymph nodes by leveraging the publicly available Camelyon16 dataset and an in-house dataset. We propose a simple yet effective self-knowledge distillation approach with contrastive embedding alignment across domains. By transferring knowledge from the source data to the target data, the experimental results demonstrated remarkable adjustments in the domain shift. These findings have significant implications for the advancement of tumor detection accuracy and reliability of histopathological WSI analysis. © 2025 Elsevier B.V., All rights reserved.

키워드