This page provides the instructions for dataset preparation on existing benchmarks, include
- Video Object Detection
- Multiple Object Tracking
- Single Object Tracking
- Video Instance Segmentation
Please download the datasets from the official websites. It is recommended to symlink the root of the datasets to $MMTRACKING/data.
-
For the training and testing of video object detection task, only ILSVRC dataset is needed.
-
The
ListsunderILSVRCcontains the txt files from here.
- For the training and testing of multi object tracking task, one of the MOT Challenge datasets (e.g. MOT17) is needed, and CrowdHuman can be served as comlementary dataset.
-
For the training and testing of single object tracking task, the MSCOCO, ILSVRC, LaSOT, UAV123, TrackingNet, OTB100, GOT10k and VOT2018 datasets are needed.
-
For OTB100 dataset, you don't need to download the dataset from the official website manually, since we provide a script to download it.
# download OTB100 dataset by web crawling
python ./tools/convert_datasets/otb100/download_otb100.py -o ./data/otb100/zips -p 8- For VOT2018, we use the official downloading script.
# download VOT2018 dataset by web crawling
python ./tools/convert_datasets/vot/download_vot.py --dataset vot2018 --save_path ./data/vot2018/data- For the training and testing of video instance segmetatioon task, only one of YouTube-VIS datasets (e.g. YouTube-VIS 2019) is needed.
If your folder structure is different from the following, you may need to change the corresponding paths in config files.
mmtracking
├── mmtrack
├── tools
├── configs
├── data
│ ├── coco
│ │ ├── train2017
│ │ ├── val2017
│ │ ├── test2017
│ │ ├── annotations
│ │
│ ├── ILSVRC
│ │ ├── Data
│ │ │ ├── DET
| │ │ │ ├── train
| │ │ │ ├── val
| │ │ │ ├── test
│ │ │ ├── VID
| │ │ │ ├── train
| │ │ │ ├── val
| │ │ │ ├── test
│ │ ├── Annotations
│ │ │ ├── DET
| │ │ │ ├── train
| │ │ │ ├── val
│ │ │ ├── VID
| │ │ │ ├── train
| │ │ │ ├── val
│ │ ├── Lists
│ │
| ├── MOT15/MOT16/MOT17/MOT20
| | ├── train
| | ├── test
│ │
│ ├── crowdhuman
│ │ ├── annotation_train.odgt
│ │ ├── annotation_val.odgt
│ │ ├── train
│ │ │ ├── Images
│ │ │ ├── CrowdHuman_train01.zip
│ │ │ ├── CrowdHuman_train02.zip
│ │ │ ├── CrowdHuman_train03.zip
│ │ ├── val
│ │ │ ├── Images
│ │ │ ├── CrowdHuman_val.zip
│ │
│ ├── lasot
│ │ ├── LaSOTBenchmark
│ │ │ ├── airplane
| │ │ │ ├── airplane-1
| │ │ │ ├── airplane-2
| │ │ │ ├── ......
│ │ │ ├── ......
│ │
│ ├── UAV123
│ │ ├── data_seq
│ │ │ ├── UAV123
│ │ │ │ ├── bike1
│ │ │ │ ├── boat1
│ │ │ │ ├── ......
│ │ ├── anno
│ │ │ ├── UAV123
│ │
│ ├── trackingnet
│ │ ├── TEST.zip
│ │ ├── TRAIN_0.zip
│ │ ├── ......
│ │ ├── TRAIN_11.zip
│ │
│ ├── otb100
│ │ │── zips
│ │ │ │── Basketball.zip
│ │ │ │── Biker.zip
│ │ │ │──
│ │
│ ├── got10k
│ │ │── full_data
│ │ │ │── train_data
│ │ │ │ ├── GOT-10k_Train_split_01.zip
│ │ │ │ ├── ......
│ │ │ │ ├── GOT-10k_Train_split_19.zip
│ │ │ │ ├── list.txt
│ │ │ │── test_data.zip
│ │ │ │── val_data.zip
│ │
| ├── vot2018
| | ├── data
| | | ├── ants1
| │ │ │ ├──color
│ │
│ ├── youtube_vis_2019
│ │ │── train
│ │ │ │── JPEGImages
│ │ │ │── ......
│ │ │── valid
│ │ │ │── JPEGImages
│ │ │ │── ......
│ │ │── test
│ │ │ │── JPEGImages
│ │ │ │── ......
│ │ │── train.json (the official annotation files)
│ │ │── valid.json (the official annotation files)
│ │ │── test.json (the official annotation files)
│ │
│ ├── youtube_vis_2021
│ │ │── train
│ │ │ │── JPEGImages
│ │ │ │── instances.json (the official annotation files)
│ │ │ │── ......
│ │ │── valid
│ │ │ │── JPEGImages
│ │ │ │── instances.json (the official annotation files)
│ │ │ │── ......
│ │ │── test
│ │ │ │── JPEGImages
│ │ │ │── instances.json (the official annotation files)
│ │ │ │── ......
We use CocoVID to maintain all datasets in this codebase. In this case, you need to convert the official annotations to this style. We provide scripts and the usages are as following:
# ImageNet DET
python ./tools/convert_datasets/ilsvrc/imagenet2coco_det.py -i ./data/ILSVRC -o ./data/ILSVRC/annotations
# ImageNet VID
python ./tools/convert_datasets/ilsvrc/imagenet2coco_vid.py -i ./data/ILSVRC -o ./data/ILSVRC/annotations
# MOT17
# The processing of other MOT Challenge dataset is the same as MOT17
python ./tools/convert_datasets/mot/mot2coco.py -i ./data/MOT17/ -o ./data/MOT17/annotations --split-train --convert-det
python ./tools/convert_datasets/mot/mot2reid.py -i ./data/MOT17/ -o ./data/MOT17/reid --val-split 0.2 --vis-threshold 0.3
# CrowdHuman
python ./tools/convert_datasets/mot/crowdhuman2coco.py -i ./data/crowdhuman -o ./data/crowdhuman/annotations
# LaSOT
python ./tools/convert_datasets/lasot/lasot2coco.py -i ./data/lasot/LaSOTBenchmark -o ./data/lasot/annotations
# UAV123
python ./tools/convert_datasets/uav123/uav2coco.py -i ./data/UAV123/ -o ./data/UAV123/annotations
# TrackingNet
# unzip files in 'data/trackingnet/*.zip'
bash ./tools/convert_datasets/trackingnet/unzip_trackingnet.sh ./data/trackingnet
# generate annotations
python ./tools/convert_datasets/trackingnet/trackingnet2coco.py -i ./data/trackingnet -o ./data/trackingnet/annotations
# OTB100
# unzip files in 'data/otb100/zips/*.zip'
bash ./tools/convert_datasets/otb100/unzip_otb100.sh ./data/otb100
# generate annotations
python ./tools/convert_datasets/otb100/otb2coco.py -i ./data/otb100 -o ./data/otb100/annotations
# GOT10k
# unzip 'data/got10k/full_data/test_data.zip', 'data/got10k/full_data/val_data.zip' and files in 'data/got10k/full_data/train_data/*.zip'
bash ./tools/convert_datasets/got10k/unzip_got10k.sh ./data/got10k
# generate annotations
python ./tools/convert_datasets/got10k/got10k2coco.py -i ./data/got10k -o ./data/got10k/annotations
# VOT2018
python ./tools/convert_datasets/vot/vot2coco.py -i ./data/vot2018 -o ./data/vot2018/annotations --dataset_type vot2018
# YouTube-VIS 2019
python ./tools/convert_datasets/youtubevis/youtubevis2coco.py -i ./data/youtube_vis_2019 -o ./data/youtube_vis_2019/annotations --version 2019
# YouTube-VIS 2021
python ./tools/convert_datasets/youtubevis/youtubevis2coco.py -i ./data/youtube_vis_2021 -o ./data/youtube_vis_2021/annotations --version 2021The folder structure will be as following after your run these scripts:
mmtracking
├── mmtrack
├── tools
├── configs
├── data
│ ├── coco
│ │ ├── train2017
│ │ ├── val2017
│ │ ├── test2017
│ │ ├── annotations
│ │
│ ├── ILSVRC
│ │ ├── Data
│ │ │ ├── DET
| │ │ │ ├── train
| │ │ │ ├── val
| │ │ │ ├── test
│ │ │ ├── VID
| │ │ │ ├── train
| │ │ │ ├── val
| │ │ │ ├── test
│ │ ├── Annotations (the official annotation files)
│ │ │ ├── DET
| │ │ │ ├── train
| │ │ │ ├── val
│ │ │ ├── VID
| │ │ │ ├── train
| │ │ │ ├── val
│ │ ├── Lists
│ │ ├── annotations (the converted annotation files)
│ │
| ├── MOT15/MOT16/MOT17/MOT20
| | ├── train
| | ├── test
| | ├── annotations
| | ├── reid
│ │ │ ├── imgs
│ │ │ ├── meta
│ │
│ ├── crowdhuman
│ │ ├── annotation_train.odgt
│ │ ├── annotation_val.odgt
│ │ ├── train
│ │ │ ├── Images
│ │ │ ├── CrowdHuman_train01.zip
│ │ │ ├── CrowdHuman_train02.zip
│ │ │ ├── CrowdHuman_train03.zip
│ │ ├── val
│ │ │ ├── Images
│ │ │ ├── CrowdHuman_val.zip
│ │ ├── annotations
│ │ │ ├── crowdhuman_train.json
│ │ │ ├── crowdhuman_val.json
│ │
│ ├── lasot
│ │ ├── LaSOTBenchmark
│ │ │ ├── airplane
| │ │ │ ├── airplane-1
| │ │ │ ├── airplane-2
| │ │ │ ├── ......
│ │ │ ├── ......
│ │ ├── annotations
│ │
│ ├── UAV123
│ │ ├── data_seq
│ │ │ ├── UAV123
│ │ │ │ ├── bike1
│ │ │ │ ├── boat1
│ │ │ │ ├── ......
│ │ ├── anno (the official annotation files)
│ │ │ ├── UAV123
│ │ ├── annotations (the converted annotation file)
│ │
│ ├── trackingnet
│ │ ├── TEST
│ │ │ ├── anno (the official annotation files)
│ │ │ ├── zips
│ │ │ ├── frames (the unzipped folders)
│ │ │ │ ├── 0-6LB4FqxoE_0
│ │ │ │ ├── 07Ysk1C0ZX0_0
│ │ │ │ ├── ......
│ │ ├── TRAIN_0
│ │ │ ├── anno (the official annotation files)
│ │ │ ├── zips
│ │ │ ├── frames (the unzipped folders)
│ │ │ │ ├── -3TIfnTSM6c_2
│ │ │ │ ├── a1qoB1eERn0_0
│ │ │ │ ├── ......
│ │ ├── ......
│ │ ├── TRAIN_11
│ │ ├── annotations (the converted annotation file)
│ │
│ ├── otb100
│ │ ├── zips
│ │ │ ├── Basketball.zip
│ │ │ ├── Biker.zip
│ │ │ │── ......
│ │ ├── annotations
│ │ ├── data
│ │ │ ├── Basketball
│ │ │ │ ├── img
│ │ │ ├── ......
│ │
│ ├── got10k
│ │ │── full_data
│ │ │ │── train_data
│ │ │ │ ├── GOT-10k_Train_split_01.zip
│ │ │ │ ├── ......
│ │ │ │ ├── GOT-10k_Train_split_19.zip
│ │ │ │ ├── list.txt
│ │ │ │── test_data.zip
│ │ │ │── val_data.zip
│ │ │── train
│ │ │ ├── GOT-10k_Train_000001
│ │ │ │ ├── ......
│ │ │ ├── GOT-10k_Train_009335
│ │ │ ├── list.txt
│ │ │── test
│ │ │ ├── GOT-10k_Test_000001
│ │ │ │ ├── ......
│ │ │ ├── GOT-10k_Test_000180
│ │ │ ├── list.txt
│ │ │── val
│ │ │ ├── GOT-10k_Val_000001
│ │ │ │ ├── ......
│ │ │ ├── GOT-10k_Val_000180
│ │ │ ├── list.txt
│ │ │── annotations
│ │
| ├── vot2018
| | ├── data
| | | ├── ants1
| │ │ │ ├──color
| | ├── annotations
│ │ │ ├── ......
│ │
│ ├── youtube_vis_2019
│ │ │── train
│ │ │ │── JPEGImages
│ │ │ │── ......
│ │ │── valid
│ │ │ │── JPEGImages
│ │ │ │── ......
│ │ │── test
│ │ │ │── JPEGImages
│ │ │ │── ......
│ │ │── train.json (the official annotation files)
│ │ │── valid.json (the official annotation files)
│ │ │── test.json (the official annotation files)
│ │ │── annotations (the converted annotation file)
│ │
│ ├── youtube_vis_2021
│ │ │── train
│ │ │ │── JPEGImages
│ │ │ │── instances.json (the official annotation files)
│ │ │ │── ......
│ │ │── valid
│ │ │ │── JPEGImages
│ │ │ │── instances.json (the official annotation files)
│ │ │ │── ......
│ │ │── test
│ │ │ │── JPEGImages
│ │ │ │── instances.json (the official annotation files)
│ │ │ │── ......
│ │ │── annotations (the converted annotation file)
There are 3 JSON files in data/ILSVRC/annotations:
imagenet_det_30plus1cls.json: JSON file containing the annotations information of the training set in ImageNet DET dataset. The 30 in 30plus1cls denotes the overlapped 30 categories in ImageNet VID dataset, and the 1cls means we take the other 170 categories in ImageNet DET dataset as a category, named as other_categeries.
imagenet_vid_train.json: JSON file containing the annotations information of the training set in ImageNet VID dataset.
imagenet_vid_val.json: JSON file containing the annotations information of the validation set in ImageNet VID dataset.
We take MOT17 dataset as examples, the other datasets share similar structure.
There are 8 JSON files in data/MOT17/annotations:
train_cocoformat.json: JSON file containing the annotations information of the training set in MOT17 dataset.
train_detections.pkl: Pickle file containing the public detections of the training set in MOT17 dataset.
test_cocoformat.json: JSON file containing the annotations information of the testing set in MOT17 dataset.
test_detections.pkl: Pickle file containing the public detections of the testing set in MOT17 dataset.
half-train_cocoformat.json, half-train_detections.pkl, half-val_cocoformat.jsonand half-val_detections.pkl share similar meaning with train_cocoformat.json and train_detections.pkl. The half means we split each video in the training set into half. The first half videos are denoted as half-train set, and the second half videos are denoted ashalf-val set.
The structure of data/MOT17/reid is as follows:
reid
├── imgs
│ ├── MOT17-02-FRCNN_000002
│ │ ├── 000000.jpg
│ │ ├── 000001.jpg
│ │ ├── ...
│ ├── MOT17-02-FRCNN_000003
│ │ ├── 000000.jpg
│ │ ├── 000001.jpg
│ │ ├── ...
├── meta
│ ├── train_80.txt
│ ├── val_20.txt
The 80 in train_80.txt means the proportion of the training dataset to the whole ReID dataset is 80%. While the proportion of the validation dataset is 20%.
For training, we provide a annotation list train_80.txt. Each line of the list contains a filename and its corresponding ground-truth labels. The format is as follows:
MOT17-05-FRCNN_000110/000018.jpg 0
MOT17-13-FRCNN_000146/000014.jpg 1
MOT17-05-FRCNN_000088/000004.jpg 2
MOT17-02-FRCNN_000009/000081.jpg 3
MOT17-05-FRCNN_000110 denotes the 110-th person in MOT17-05-FRCNN video.
For validation, The annotation list val_20.txt remains the same as format above.
Images in reid/imgs are cropped from raw images in MOT17/train by the corresponding gt.txt. The value of ground-truth labels should fall in range [0, num_classes - 1].
There are 2 JSON files in data/crowdhuman/annotations:
crowdhuman_train.json: JSON file containing the annotations information of the training set in CrowdHuman dataset.
crowdhuman_val.json: JSON file containing the annotations information of the validation set in CrowdHuman dataset.
There are 2 JSON files in data/lasot/annotations:
lasot_train.json: JSON file containing the annotations information of the training set in LaSOT dataset.
lasot_test.json: JSON file containing the annotations information of the testing set in LaSOT dataset.
There are only 1 JSON files in data/UAV123/annotations:
uav123.json: JSON file containing the annotations information of the UAV123 dataset.
There are 511 video directories of TrackingNet testset in data/trackingnet/TEST/frames, and each video directory contains all images of the video. Similar file structures can be seen in data/trackingnet/TRAIN_{*}/frames.
There are 2 JSON files in data/trackingnet/annotations:
trackingnet_test.json: JSON file containing the annotations information of the testing set in TrackingNet dataset.
trackingnet_train.json: JSON file containing the annotations information of the training set in TrackingNet dataset.
There are 98 video directories of OTB100 dataset in data/otb100/data, and the img folder under each video directory contains all images of the video.
There are only 1 JSON files in data/otb100/annotations:
otb100.json: JSON file containing the annotations information of the OTB100 dataset.
There are training video directories in data/got10k/train, and each video directory contains all images of the video. Similar file structures can be seen in data/got10k/test and data/got10k/val.
There are 3 JSON files in data/got10k/annotations:
got10k_train.json: JSON file containing the annotations information of the training set in GOT10k dataset.
got10k_test.json: JSON file containing the annotations information of the testing set in GOT10k dataset.
got10k_val.json: JSON file containing the annotations information of the valuation set in GOT10k dataset.
There are 60 video directories of VOT2018 dataset in data/vot2018/data, and the color folder under each video directory contains all images of the video.
There are only 1 JSON files in data/vot2018/annotations:
vot2018.json: JSON file containing the annotations information of the VOT2018 dataset.
There are 3 JSON files in data/youtube_vis_2019/annotations or data/youtube_vis_2021/annotations:
youtube_vis_2019_train.json/youtube_vis_2021_train.json: JSON file containing the annotations information of the training set in youtube_vis_2019/youtube_vis2021 dataset.
youtube_vis_2019_valid.json/youtube_vis_2021_valid.json: JSON file containing the annotations information of the validation set in youtube_vis_2019/youtube_vis2021 dataset.
youtube_vis_2019_test.json/youtube_vis_2021_test.json: JSON file containing the annotations information of the testing set in youtube_vis_2019/youtube_vis2021 dataset.