CycleGAN Generated Synthetic Dataset
收藏arXiv2021-07-15 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/2107.07271v1
下载链接
链接失效反馈官方服务：
资源简介：
本研究介绍了名为CycleGAN Generated Synthetic Dataset的数据集，由女王大学电子、电气工程与计算机科学学院的研究团队创建。该数据集包含20,000个组织图像三元组，每个三元组代表同一组织结构但具有不同的染色外观。数据集通过使用CycleGAN模型，将来自不同染色和扫描条件的数据映射到新的已知域中，以评估计算模型在多域应用中的性能。此数据集的创建旨在解决数字病理学中因不同染色和扫描协议导致的域移问题，提高机器学习模型在自动化组织病理学管道中的性能。
This study introduces a dataset named CycleGAN Generated Synthetic Dataset, which was developed by the research team from the School of Electronics, Electrical Engineering and Computer Science at Queen's University. The dataset contains 20,000 tissue image triplets, where each triplet represents the same tissue structure but with distinct staining appearances. It uses the CycleGAN model to map data from different staining and scanning conditions into new, known domains, so as to evaluate the performance of computational models in multi-domain applications. This dataset was created to address the domain shift issue caused by varying staining and scanning protocols in digital pathology, and to improve the performance of machine learning models in automated histopathology pipelines.
提供机构：
女王大学电子、电气工程与计算机科学学院，贝尔法斯特，北爱尔兰，英国
创建时间：
2021-07-15
