Medical Segmentation Decathlon (MSD) collection
收藏资源简介:
COLosSAL是一个针对3D医学图像分割的冷启动主动学习基准,基于公开的医学分割十项全能(MSD)数据集构建。该数据集包含五项任务,涵盖CT和MRI两种常见3D图像模式,用于健康组织和肿瘤/病理的分割。数据集的创建过程涉及对未标记数据池的样本选择,旨在通过最小化标注样本数量来提高标注效率。COLosSAL的应用领域主要集中在医学图像分析,特别是3D医学图像的分割任务,旨在解决数据标注过程中的瓶颈问题,提高深度学习模型的训练效率。
COLosSAL is a cold-start active learning benchmark for 3D medical image segmentation, built upon the publicly available Medical Segmentation Decathlon (MSD) dataset. This benchmark includes five tasks, covering two common 3D imaging modalities: CT and MRI, which are designed for the segmentation of healthy tissues, tumors and pathologies. The construction of COLosSAL involves sample selection from the unlabeled data pool, aiming to improve annotation efficiency by minimizing the number of labeled samples. The primary application domains of COLosSAL focus on medical image analysis, especially 3D medical image segmentation tasks, and it is intended to address the bottleneck issues in the data annotation process and enhance the training efficiency of deep learning models.

- 1COLosSAL: A Benchmark for Cold-start Active Learning for 3D Medical Image Segmentation范德堡大学 · 2023年



