Skip to content

Latest commit

 

History

History
67 lines (59 loc) · 2.96 KB

DATASET.md

File metadata and controls

67 lines (59 loc) · 2.96 KB
Raw

Dataset

Disclaimer

This software is for non-commercial use only. The source code is released under the Attribution-NonCommercial-ShareAlike (CC BY-NC-SA) Licence. Permission is granted to use the data given that you agree to both our license terms and that of the original Cityscapes-dataset

Download datasets

a. Symlink the $DATA_ROOT dataset to $MMDETECTION/data folder.

b. Download Cityscapes-VPS here in $CITY_VPS = data/cityscapes_vps/ folder.

c. Download leftImg8bit_sequence.zip and gtFine.zip from the Cityscapes-dataset webpage in data folder. You need only val/ of these datasets to construct Cityscapes-VPS.

Merge Cityscapes into Cityscapes-VPS

We have 2400/600/600 frames for train/val/test splits. Fetch cityscapes sequence images into $CITY_VPS/SPLIT/img and $CITY_VPS/SPLIT/img_all, and merge two datasets labels at $CITY_VPS/SPLIT/cls and $CITY_VPS/SPLIT/inst.

bash ./prepare_data/merge_datasets.sh \
    data/cityscapes_vps  data/leftImg8bit_sequence/val/  data/gtFine/val/
# OR
# For SPLIT = 'train'/'val'/'test'
python prepare_data/fetch_city_images.py \
    --src_dir data/leftImg8bit_sequence/val/ \
    --dst_dir $CITY_VPS --mode SPLIT 
python prepare_data/merge_datasets.py \
    --src_dir data/gtFine/val/ \
    --dst_dir $CITY_VPS --mode SPLIT

Create video-panoptic labels

Create labelmap/, panoptic_inst/, and panoptic_video/ in $CITY_VPS/SPLIT/ by running following commands.

bash ./prepare_data/create_panoptic_labels.sh \
    data/cityscapes_vps/
# OR
# For SPLIT = 'train'/'val'
python prepare_data/create_panoptic_labels.py --root_dir $CITY_VPS --mode SPLIT
# For SPLIT = 'val'
python prepare_data/create_panoptic_video_labels.py --root_dir $CITY_VPS --mode SPLIT

Directory Structure

Necessary data for training, testing, and evaluation are as follows. You may delete other data for disk usage.

mmdetection
├── mmdet
├── tools
├── configs
├── data
│   ├── cityscapes_vps
│   │   ├── panoptic_im_train_city_vps.json
│   │   ├── panoptic_im_val_city_vps.json
│   │   ├── panoptic_im_test_city_vps.json  
│   │   ├── instances_train_city_vps_rle.json (for training)
│   │   ├── instances_val_city_vps_rle.json 
│   │   ├── im_all_info_val_city_vps.json (for inference)
│   │   ├── im_all_info_test_city_vps.json (for inference)
│   │   ├── panoptic_gt_val_city_vps.json (for VPQ eval)
│   │   ├── train 
│   │   │   ├── img
│   │   │   ├── labelmap
│   │   ├── val
│   │   │   ├── img
│   │   │   ├── img_all
│   │   │   ├── panoptic_video
│   │   ├── test
│   │   │   ├── img_all