scripts for downloading scannet dataset and extract them using multiprocessing to speed up
Please refer to scannet official repo for downloading commands.
The official download script uses python2.7 so we will use multiprocessing library to speed up extraction of .sens files.
The downloaded data structure should be:
extract.py
reader.py
SensorData.py
download_scannet.py
download_scannetv2.py
/data_downloaded
/scans_test
/scene0707_00
scene0707_00.sens
...
/scans_train
/scene0000_00
scene0000_00.sens
...
Then modify the split(train/test) of the file you want to decompress and the path of the python interpreter to use in extract.py
Also, Specify the maxprocess variable in extract.py
After running extract.py, the file structure should be:
/data_extracted
/scans_test
/scene0707_00
/color
000000.jpg
/depth
000000.png
/pose
000000.txt
intrinsics_color.txt
intrinsics_depth.txt
...
/scans_train
...