CERVIX93
收藏资源简介:
CERVIX93数据集由南缅因大学和南佛罗里达大学共同创建,包含93个真实的宫颈细胞学图像堆栈,用于评估核检测和图像分类方法。数据集中的图像来自不同等级的ThinPrep细胞学幻灯片,共有2705个手动标记的核。创建过程中使用了SRC Biosciences的Stereologer系统进行图像采集和手动分级。该数据集主要用于评估和改进宫颈核分割和检测技术,以辅助宫颈癌的早期诊断和治疗。
The CERVIX93 dataset was jointly developed by the University of Southern Maine and the University of South Florida, comprising 93 stacks of real cervical cytology images for the evaluation of nuclear detection and image classification methodologies. The images within this dataset are derived from ThinPrep cytology slides of varying grades, and include a total of 2705 manually annotated nuclei. During the dataset construction, the Stereologer system from SRC Biosciences was employed for image acquisition and manual grading. This dataset is primarily utilized to evaluate and optimize cervical nuclear segmentation and detection technologies, thereby supporting early diagnosis and treatment of cervical cancer.




