coco-annotator

coco-annotator copied to clipboard

Here's a COCO annotation format: http://cocodataset.org/#format-data

There is a derived dataset called COCO-Text: Dataset for Text Detection and Recognition with format: https://vision.cornell.edu/se3/coco-text-2/

The annotations in COCO have format:

annotation{
    "id" : int,
    "image_id" : int,
    "category_id" : int,
    "segmentation" : RLE or [polygon],
    "area" : float,
    "bbox" : [x,y,width,height],
    "iscrowd" : 0 or 1,
}

And the COCO-Text annotation looks like this:

annotation{
    “id”                     :   int,
    “image_id”               :   int,
    “class”                  :   str     # ‘machine printed’ or ‘handwritten’ or ‘others’
    “legibility”             :   str     # ‘legible’ or ‘illegible’
    “language”               :   str     # ‘english’ or ‘not english’ or ‘na’
    “area”                   :   float,
    “bbox”                   :   [x,y,width,height],
    “utf8_string”            :   str,
    “polygon”                :   []
}

You could set this as a the metadata and then write a script that brings everything from the metadata section into the body.

@orian Did you create any script to convert from COCO or any other format to COCO Text? If yes, can you please share?