Supported Formats
List of supported formats:
ADE20k (v2017) (import-only)
ADE20k (v2020) (import-only)
Align CelebA (
classification
,landmarks
) (import-only)CamVid (
segmentation
)CelebA (
classification
,detection
,landmarks
) (import-only)CIFAR-10/100 (
classification
(python version))Cityscapes (
segmentation
)CVAT (
for images
,for video
(import-only))ICDAR13/15 (
word_recognition
,text_localization
,text_segmentation
)ImageNet (
classification
,detection
)Detection format is the same as in PASCAL VOC
KITTI (
segmentation
,detection
)KITTI 3D (
raw
/tracklets
/velodyne points
)LabelMe (
labels
,boxes
,masks
)LFW (
classification
,person re-identification
,landmarks
)Mapillary Vistas (import-only)
Market-1501 (
person re-identification
)MARS (import-only)
MNIST (
classification
)MNIST in CSV (
classification
)MOT sequences
MOTS (png)
MPII Human Pose Dataset (
detection
,pose estimation
) (import-only)MPII Human Pose Dataset (JSON) (
detection
,pose estimation
) (import-only)MS COCO (
image_info
,instances
,person_keypoints
,captions
,labels
,panoptic
,stuff
)labels
are our extension - likeinstances
with onlycategory_id
Open Images (
classification
,detection
,segmentation
)PASCAL VOC (
classification
,detection
,segmentation
(class, instances),action_classification
,person_layout
)Supervisely (
pointcloud
)SYNTHIA (
segmentation
) (import-only)TF Detection API (
bboxes
,masks
)VGGFace2 (
landmarks
,bboxes
)VoTT CSV (
detection
) (import-only)VoTT JSON (
detection
) (import-only)WIDER Face (
bboxes
)YOLO (
bboxes
)
Supported annotation types
Labels
Bounding boxes
Polygons
Polylines
(Segmentation) Masks
(Key-)Points
Captions
3D cuboids
Datumaro does not separate datasets by tasks like classification, detection etc. Instead, datasets can have any annotations. When a dataset is exported in a specific format, only relevant annotations are exported.
Dataset meta info file
It is possible to use classes that are not original to the format.
To do this, use dataset_meta.json
.
{
"label_map": {"0": "background", "1": "car", "2": "person"},
"segmentation_colors": [[0, 0, 0], [255, 0, 0], [0, 0, 255]],
"background_label": "0"
}
label_map
is a dictionary where the class ID is the key and the class name is the value.segmentation_colors
is a list of channel-wise values for each class. This is only necessary for the segmentation task.background_label
is a background label ID in the dataset.