NuInsSeg
收藏资源简介:
NuInsSeg是由医学图像分析与人工智能研究中心和维也纳医科大学病理生理学与过敏研究所共同创建的大型全注释数据集,专注于H&E染色的组织图像中的细胞核实例分割。该数据集包含665个图像补丁,超过30,000个手动分割的细胞核,来自31个人类和鼠类器官。此外,首次提供了整个数据集的模糊区域掩码,这些区域代表了图像中精确和确定性手动注释不可能的部分。NuInsSeg数据集旨在通过提供高质量的训练数据,推动深度学习模型在医学图像分析中的应用,特别是在细胞核分割任务中,以提高疾病诊断的准确性和效率。
NuInsSeg is a large-scale fully annotated dataset jointly developed by the Medical Image Analysis and Artificial Intelligence Research Center and the Institute of Pathophysiology and Allergology, Medical University of Vienna. It focuses on nuclear instance segmentation in H&E-stained histological images. The dataset comprises 665 image patches with over 30,000 manually segmented cell nuclei, sourced from 31 human and rodent organs. Additionally, fuzzy region masks covering the entire dataset are provided for the first time, which represent areas in the images where precise and definitive manual annotation is infeasible. The NuInsSeg dataset aims to advance the application of deep learning models in medical image analysis, particularly for nuclear segmentation tasks, by providing high-quality training data to improve the accuracy and efficiency of disease diagnosis.

- 1NuInsSeg: A Fully Annotated Dataset for Nuclei Instance Segmentation in H&E-Stained Histological Images医学图像分析与人工智能研究中心 · 2023年



