EGP3D Dataset
收藏资源简介:
EGP3D Dataset是由香港中文大学(深圳)数据科学学院等机构创建的,专门用于RGB-D相机点云超分辨率任务的数据集。该数据集包含了从简单几何形状到复杂人像和恐龙模型的多种对象,每个模型从多个角度捕获,确保数据的全面性和复杂性。与现有数据集相比,EGP3D Dataset更真实地反映了现实世界中的点云特征,包括噪声、杂散光等挑战因素。数据集的创建过程结合了RGB图像的边缘信息,通过几何优化方法生成高质量的点云数据。该数据集主要应用于3D重建和机器人导航等领域,旨在解决低分辨率点云在实际应用中的不足。
The EGP3D Dataset was developed by the School of Data Science, The Chinese University of Hong Kong, Shenzhen and other institutions, and is specifically designed for the point cloud super-resolution task using RGB-D cameras. It encompasses a wide range of objects spanning from simple geometric shapes to complex human and dinosaur models, with each model captured from multiple viewpoints to ensure the comprehensiveness and complexity of the data. Compared with existing datasets, the EGP3D Dataset more authentically reflects real-world point cloud characteristics, including challenging factors such as noise and stray light. The dataset was created by incorporating edge information from RGB images and generating high-quality point cloud data through geometric optimization methods. This dataset is mainly applied in fields such as 3D reconstruction and robotic navigation, aiming to address the shortcomings of low-resolution point clouds in practical applications.




