Closed YunzeMan closed 1 year ago
Take a look at the available subset lists. All redacted samples are stored in respective test sets, while the unredacted ones are in validation/train sets.
Hence, make sure to load the correct set in order to avoid loading the redacted images.
I notice that some sequences in the dataset don't have GT depth map (held out deliberately for testing)
However, I'd like to use this dataset for depth estimation. So I was wondering how I can filter the sequences with depth maps and camera parameters available and remove those without depth.
Thanks!