We train our model on the VLOG Dataset. We use the official release of the videos, the files are named as "block_x.tar" (0<=x<=4). We assume the videos are downloaded on the path: YOUR_DATASET_FOLDER/vlog/.
Download the list for the videos data_v1.1.tgz. Extract the list "manifest.txt" to the same folder: YOUR_DATASET_FOLDER.
Go into the folder:
cd preprocessChange the video path in preprocess/downscale_video_joblib.py. Reduce the video size and save it to YOUR_DATASET_FOLDER/vlog_256/ :
python downscale_video_joblib.pyExtract the jpgs to YOUR_DATASET_FOLDER/vlog_frames_12fps/ by using:
python extract_jpegs_256.pyGnerate the jpg list to YOUR_DATASET_FOLDER/vlog_frames_12fps.txt for training:
python genvloglist.py- DAVIS2017
Download DAVIS 2017 dataset on the path: YOUR_DATASET_FOLDER/davis/ . Clone the evaluation code for DAVIS 2017 to YOUR_DATASET_FOLDER/davis-2017/ .
Go into the folder:
cd preprocessGenerate the groundtruth and the list for testing as YOUR_DATASET_FOLDER/davis/DAVIS/strip_vallist.txt of the texture propagation task:
python genstripe.py -d YOUR_DATASET_FOLDER/davis/Generate the list for instance segmentation testing as YOUR_DATASET_FOLDER/davis/DAVIS/vallist.txt:
python gendavis_vallist.py -d YOUR_DATASET_FOLDER/davis/-
JHMDB Download JHMDB dataset (joint_ositions.zip, puppet_mask.zip, Rename_images.tar.gz, splits.zip) and extract them on the path YOUR_DATSET_FOLDER/JHMDB/.
./JHMDB ├── joint_positions │ └──├── brush_hair │ ├── catch │ ├── clap │ ├── climb_stairs │ ├── golf │ ├── ... │ └── wave ├── puppet_mask │ └── ........... same as joint_positions ├── Rename_Images │ └── ........... same as joint_positions └── splits └──├── brush_hair_test_split1.txt ├── brush_hair_test_split2.txt ├── brush_hair_test_split3.txt ├── catch_test_split1.txt ├── ... └── wave_test_split3.txt
Go into the folder:
cd preprocessGenerate the list for testing as YOUR_DATASET_FOLDER/JHMDB/vallist.txt:
python genJHMDB_vallist.py -d YOUR_DATASET_FOLDER/JHMDB/