Dishwasher30k
收藏arXiv2023-09-21 更新2024-06-21 收录
下载链接：
https://florise.github.io/neural labeling web/
下载链接
链接失效反馈官方服务：
资源简介：
Dishwasher30k数据集由国家先进工业科学技术研究所创建，包含30000帧透明物体的RGB和噪声深度图像，这些图像是在洗碗机中使用RGBD传感器捕获的。数据集创建过程中，通过NeuralLabeling工具添加了地面真实深度图，以提高深度重建性能。该数据集主要应用于机器人领域，旨在解决透明物体深度完成的问题，通过监督学习提高深度估计的准确性。
The Dishwasher30k dataset was created by the National Institute of Advanced Industrial Science and Technology (AIST). It contains 30,000 frames of RGB and noisy depth images of transparent objects, which were captured using an RGBD sensor inside a dishwasher. During the dataset construction process, ground-truth depth maps were added via the NeuralLabeling tool to enhance depth reconstruction performance. This dataset is primarily applied in the robotics domain, with the objective of solving the depth completion problem for transparent objects and improving the accuracy of depth estimation through supervised learning.
提供机构：
国家先进工业科学技术研究所
创建时间：
2023-09-21
搜集汇总
数据集介绍
以上内容由遇见数据集搜集并总结生成
