Huang Wentao, Hu Xiaoling, Abousamra Shahira, Prasanna Prateek, Chen Chao
Stony Brook University, Stony Brook, NY, USA.
Harvard Medical School, Boston, MA, USA.
Med Image Comput Comput Assist Interv. 2024 Oct;15004:144-154. doi: 10.1007/978-3-031-72083-3_14. Epub 2024 Oct 14.
Weakly supervised whole slide image (WSI) classification is challenging due to the lack of patch-level labels and high computational costs. State-of-the-art methods use self-supervised patch-wise feature representations for multiple instance learning (MIL). Recently, methods have been proposed to fine-tune the feature representation on the downstream task using pseudo labeling, but mostly focusing on selecting high-quality positive patches. In this paper, we propose to mine hard negative samples during fine-tuning. This allows us to obtain better feature representations and reduce the training cost. Furthermore, we propose a novel patch-wise ranking loss in MIL to better exploit these hard negative samples. Experiments on two public datasets demonstrate the efficacy of these proposed ideas. Our codes are available at https://github.com/winston52/HNM-WSI.
由于缺乏切片级标签和高计算成本，弱监督全切片图像（WSI）分类具有挑战性。当前的先进方法使用自监督的切片级特征表示进行多实例学习（MIL）。最近，已经有人提出使用伪标签在下游任务上微调特征表示的方法，但大多集中在选择高质量的正样本切片上。在本文中，我们建议在微调过程中挖掘硬负样本。这使我们能够获得更好的特征表示并降低训练成本。此外，我们在MIL中提出了一种新颖的切片级排序损失，以更好地利用这些硬负样本。在两个公共数据集上进行的实验证明了这些提出的想法的有效性。我们的代码可在https://github.com/winston52/HNM-WSI上获取。
