datasets
datasets copied to clipboard
pose headers
So i am looking through the datasets and some have just a pose url, whereas others also have a pose header. It is my understanding that the datasets with a pose header || pose header file are the ones that can create or also have the .pose files? Can you confirm this?
All datasets where a pose header is available, have an external download of pose files. The pose header is used when reading the dataset after it is saved to disk. Since for effieciency, on disk, the poses are stored as numpy arrays, they miss that header information.