HC Depth数据集
收藏arXiv2020-07-22 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/2007.11256v1
下载链接
链接失效反馈官方服务：
资源简介：
HC Depth数据集是由快手科技创建的，旨在解决现有单目深度估计方法在特殊光照条件、动态物体和倾斜相机角度等挑战性场景下的不足。该数据集包含120,060张图像，其中24,660张使用微软Kinect采集，主要覆盖室内人像场景；95,400张使用英特尔RealSense采集，覆盖室内外场景。数据集通过语义分割和表面平滑处理，增强了深度图的质量。HC Depth数据集的应用领域包括3D重建、场景理解和增强现实，旨在提高单目深度估计的准确性和泛化能力。
HC Depth Dataset was developed by Kuaishou Technology to address the limitations of existing monocular depth estimation methods in challenging scenarios including special lighting conditions, dynamic objects, and tilted camera angles. This dataset consists of 120,060 images, of which 24,660 were captured using Microsoft Kinect and primarily cover indoor portrait scenes, while the remaining 95,400 images were collected using Intel RealSense, covering both indoor and outdoor scenes. The quality of depth maps in this dataset is enhanced via semantic segmentation and surface smoothing processing. Applications of the HC Depth Dataset span 3D reconstruction, scene understanding and augmented reality, aiming to improve the accuracy and generalization ability of monocular depth estimation models.
提供机构：
快手科技
创建时间：
2020-07-22
