Prepare NYU Depth V2 dataset

@inproceedings{Silberman:ECCV12,
  author    = {Nathan Silberman, Derek Hoiem, Pushmeet Kohli and Rob Fergus},
  title     = {Indoor Segmentation and Support Inference from RGBD Images},
  booktitle = {ECCV},
  year      = {2012}
}

This project use the preprocessed NYU Depth V2 dataset in HDF5 formats. You can download here. This dataset requires 32G of storage space.

You can download the dataset by running the following command in project root directory:

sh data_prepare/NYU_Depth/download_nyu.sh

and the folder structure should be like this:

data
|   nyudepth_hdf5
|   |   |   ├── train
|   |   |   |   ├── basement_0001a
|   |   |   |   |   ├── xxxx.h5
|   |   |   |   ├── basement_0001b
|   |   |   |   |   ├── xxxx.h5
|   |   |   ├── val
|   |   |   |   ├── official
|   |   |   |   |   ├── xxxx.h5
|   |   |   ├── train.csv
|   |   |   ├── val.csv