期刊
COMPUTER VISION, ECCV 2022, PT XXI
卷 13681, 期 -, 页码 699-715出版社
SPRINGER INTERNATIONAL PUBLISHING AG
DOI: 10.1007/978-3-031-19803-8_41
关键词
Whole-slide image classification; Multiple instance learning; Multi-scale zooming; Efficient computational pathology
This paper introduces a method called ZoomMIL, which achieves high performance in classifying gigapixel-sized Whole-Slide Images (WSIs) in digital pathology. The method builds WSI representations by aggregating tissue-context information from multiple magnifications, significantly reducing computational demands.
Multiple Instance Learning (MIL) methods have become increasingly popular for classifying gigapixel-sized Whole-Slide Images (WSIs) in digital pathology. Most MIL methods operate at a single WSI magnification, by processing all the tissue patches. Such a formulation induces high computational requirements and constrains the contextualization of the WSI-level representation to a single scale. Certain MIL methods extend to multiple scales, but they are computationally more demanding. In this paper, inspired by the pathological diagnostic process, we propose ZoomMIL, a method that learns to perform multi-level zooming in an end-to-end manner. ZoomMIL builds WSI representations by aggregating tissue-context information from multiple magnifications. The proposed method outperforms the state-of-the-art MIL methods in WSI classification on two large datasets, while significantly reducing computational demands with regard to Floating-Point Operations (FLOPs) and processing time by 40-50x. Our code is available at: https://github. com/histocartography/zoommil.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据