Units for depth and translation

[xuhaozheng]

Hi There,

Thank you for the dataset and repo. I found a problem when I tried to use the depth information and camera position together. When I checked the MOVi-F dataset, I found the depth map range is between 0 and 50, while the position such as [-1.36136764 -0.94866489 0.53709911] should be in meter. I tried to re-project the image1 from first frame view to image view 2, given the image1, depth map1, intrinsic1 and RT1 RT2. Here's the reproject results:

I wonder whether the scale of depth and that of depth map is consistent. If not, is there any way to normalize tham?