Exporting Annotations

How Do I Export Annotations?

  1. Go to the Annotations tab - You can find the tab on the top of the Datasets homepage.

  2. Select Export Annotations - On selection of the Export Annotations button, you will be greeted with the following popup with the options list below. You should select your preferred settings based on the explanations in the table below.

  3. Select Export - The process will lead to an export being created in the Current Jobs section at the bottom.

  4. Select Download - This will download the exported annotations in a zipped folder. The exported annotation link will expire, so be sure to download it when it becomes available.


Export Options

The following tables outline the different export options and details.

Options ListInput ChoicesDescription
NormalizedOn/OffA flag denoting if the annotations should be normalized with the following formula (bbox_y / height), (bbox_x / width)
ShuffleOn/OffA flag denoting if the images should be shuffled in the process of exporting
Train-Test SplitSlider value from 0.0-1.0A value that represents the proportion of the dataset to include in the test split, i.e. for a 0.3 train-test split ratio applied on a 100 image dataset, 30 images will be placed into test set. To get all annotations in one set rather than split between train and validate, you can set the train-test split to 0.
Export FormatSee Annotation Formats belowThe format of the exported annotation (see Uploading Annotations for more information) or the table below

Annotation Formats

Please note that each annotation format has a set output file type. Nexus supports a variety of annotation formats and we are striving to constantly cover more formats from different tools. To see more detailed information about the different Annotation Types, look through them in Supported Annotation Formats or locate them in the table below.

Bounding Boxes

Annotation TypeDescriptionRequired File Type
COCOThis file format is typically export from COCO Annotator and LabelMe. The sample expected file format that is accepted is as followsJSON
CSV Four CornersThis is another typical format exported when accessing Kaggle Datasets. The file is presented in a .csv format where individual rows represent 1 annotation each and the headers must be as suchCSV
CSV Width HeightAnother common CSV based representation of image annotation where the width and height of the bounding box is givenCSV
Pascal VOCThe PascalVOC annotation type is commonly exported from LabelImg and should be in an xml filetype.XML
YOLO DarkNetThis annotation format is commonly prepared to train YOLO models and it contains a label file and multiple .txt files for describing each image's annotations.TXT
YOLO Keras PyTorchThis format is almost the same as the YOLO format above, however it allows a single .txt file to describe all the annotations as suchTXT
CreateMLThis file format is typically exported from CreateML which has items per image.JSON
TFRecordThis file format is the TFRecord. More information can be found on TensorFlow.TFRecord

Polygons/Masks

Annotation TypeDescriptionRequired File Type
TFRecord Polygon / MasksThis file format is the TFRecord. More information can be found on TensorFlow.TFRecord
COCO Annotator Polygons / MasksThis uses the same COCO JSON format except the annotations component should be replaced by something like the following example below.JSON
LabelMe Mask / PolygonThis file format is typically exported from LabelMe and provides 1 annotation file per image. Users can upload every individual annotation file that looks as such:JSON

Classification

Annotation TypeDescriptionRequired File Type
TFRecord ClassificationThis file format is the TFRecord. More information can be found on TensorFlow.TFRecord
CSV ClassificationThis is simply a CSV list mapping each image to a class label.CSV

Keypoints

Annotation TypeDescriptionRequired File Type
TFRecord KeypointsThis file format is the TFRecord. More information can be found on TensorFlow.TFRecord
COCO KeypointSimilar to the original COCO format, but instead of having the segmentation vertices, you contain keypoints as a list of arrays.JSON

Additionally, we include a file for predefined keypoint skeletons on export using our Datature Skeleton format. This is to make pre-existing skeletons compatible with our skeleton editor.

Skeleton File TypeDescriptionRequired File Type
Datature SkeletonThis is our own custom skeleton schema, that integrates anything that would be needed.JSON