Endoscopy artifact detection (EAD 2019) challenge dataset
收藏资源简介:
EAD 2019挑战数据集是由牛津大学等六个国际机构合作创建的大型多模态内窥镜视频数据集,旨在解决内窥镜视频中常见的多种图像伪影问题,如像素饱和、运动模糊等。数据集包含2147个标注帧,覆盖多种组织类型和成像模式,通过随机混合收集的数据确保了多样性。创建过程中,专家临床医生进行了详细的标注,确保了数据的质量和准确性。该数据集的应用领域广泛,包括但不限于提高内窥镜视频的质量,辅助疾病诊断和治疗,以及支持计算机辅助工具的开发,从而提升患者护理质量。
The EAD 2019 Challenge Dataset is a large-scale multimodal endoscopic video dataset co-created by six international institutions including the University of Oxford. It is designed to address various common image artifacts in endoscopic videos, such as pixel saturation and motion blur. The dataset contains 2147 annotated frames covering multiple tissue types and imaging modalities, with diversity ensured by randomly mixing the collected data. During its creation, expert clinicians performed detailed annotations to guarantee the quality and accuracy of the dataset. This dataset has a wide range of applications, including but not limited to improving the quality of endoscopic videos, assisting disease diagnosis and treatment, supporting the development of computer-aided tools, and thereby enhancing patient care quality.




