Datasets:

---

binhng

/

libero_10_lerobot_mask_depth

like 0

Thank you for sharing this dataset. I am currently processing the image and depth data and would like to confirm two parameter settings to ensure accurate 3D projection:

1. Camera Intrinsics (Focal Length): For the 256x256 resolution images, does the dataset use the standard Robosuite default FOV (45°), which corresponds to a focal length of ~309.0, or does it use a different setting (e.g., f=256.0)?

2. Depth Scale Factor: What is the specific proportional coefficient (scale factor) used to convert the raw integer depth values into actual metric distance (meters)?

Thank you for your help.