Open-Source Periorbital Segmentation Dataset
收藏资源简介:
Open-Source Periorbital Segmentation Dataset是由伊利诺伊大学芝加哥分校医学院创建的,用于眼科整形和颅面分割任务的开源数据集。该数据集包含2842张图像,涵盖了虹膜、巩膜、眼睑、结膜和眉毛的详细标注。数据集的创建过程包括从两个开源数据集(Chicago Facial Dataset和CelebAMask-HQ)中提取图像,并由五名经过训练的标注者进行标注。该数据集的应用领域主要集中在眼科整形手术中的分割任务,旨在通过深度学习模型提高分割的准确性和临床实用性。
Open-Source Periorbital Segmentation Dataset was developed by the College of Medicine, University of Illinois Chicago, as an open-source dataset for ophthalmic plastic and craniofacial segmentation tasks. This dataset contains 2842 images with detailed annotations for the iris, sclera, eyelids, conjunctiva, and eyebrows. The dataset construction process involved extracting images from two existing open-source datasets, namely the Chicago Facial Dataset and CelebAMask-HQ, followed by manual annotation conducted by five trained annotators. The primary application scenarios of this dataset focus on segmentation tasks in ophthalmic plastic surgery, aiming to improve segmentation accuracy and clinical practicality through deep learning models.




