facebookresearch
Data Prepare
May I ask how depth maps are obtained when preparing datasets like re10k or DL3DV, where only camera poses are provided?
We labelled it by ourselves. Cut3r has provided their own processed ones, as can be found here, if you want to use (I did not check their quality before though)
https://github.com/CUT3R/CUT3R/blob/a0aedf8c7ff2a46c98added83c4d331ce9cd4d50/docs/preprocess.md
Hi~ regarding labelling "re10k" and "DL3DV" by yourselves, Could you please specify how these labels were generated? Are they manually annotated, or are they derived from a process like COLMAP (or another automated method)?
Thx
Hi~ regarding labelling "re10k" and "DL3DV" by yourselves, Could you please specify how these labels were generated? Are they manually annotated, or are they derived from a process like COLMAP (or another automated method)?
@jytime My dataset includes LiDAR and camera poses, but lacks depth GT. Is it possible to generate depth maps using models like Depth Anything or Metric3D, or do you recommend another method that leverages LiDAR information for depth map generation?
Thx
嗨~关于你们自己标注“re10k”和“DL3DV”，请具体说明这些标签是如何生成的吗？它们是手动注释的，还是源自 COLMAP 等流程（或其他自动化方法）？
我的数据集包括 LiDAR 和相机姿势，但缺少深度 GT。是否可以使用 Depth Anything 或 Metric3D 等模型生成深度图，或者您是否推荐另一种利用 LiDAR 信息生成深度图的方法？
感谢
Hi~ regarding labelling "re10k" and "DL3DV" by yourselves, Could you please specify how these labels were generated? Are they manually annotated, or are they derived from a process like COLMAP (or another automated method)?
@jytime My dataset includes LiDAR and camera poses, but lacks depth GT. Is it possible to generate depth maps using models like Depth Anything or Metric3D, or do you recommend another method that leverages LiDAR information for depth map generation?
Thx
Hello, thank you for your question. Have you finished generating the depth map now? What did you use to generate it?
