metadata
license: cc-by-nc-sa-4.0
task_categories:
- depth-estimation
tags:
- robotics
- transparent-objects
Status: Complete for the scope described below.
TransCG archive mirror
Original authors: Hongjie Fang, Hao-Shu Fang, Sheng Xu, and Cewu Lu. Source: https://graspnet.net/transcg and https://github.com/Galaxies99/TransCG Paper: https://doi.org/10.1109/LRA.2022.3183256
This unofficial mirror contains the 13 original data ZIPs and the metadata ZIP, covering scenes 1–130. Archives are unchanged. Extract all ZIPs into one directory. The source verification manifest records 34,191 training and 23,524 test samples. Original license: https://creativecommons.org/licenses/by-nc-sa/4.0/ Retain attribution; noncommercial use and share-alike terms apply.
manifest.json lists verified archives and SHA-256 checksums.