search

cs-chan / Total-Text-Dataset

Total Text Dataset. It consists of 1555 images with more than 3 different text orientations: Horizontal, Multi-Oriented, and Curved, one of a kind.
BSD 3-Clause "New" or "Revised" License
747 stars 140 forks source link

TotalText annotation #33

Closed yustiks closed 3 years ago

yustiks commented 3 years ago

Hello! I try to understand format of input file. Can you please explain what every field in the following format means? {"image_id": 8, "bbox": [551.0, 193.0, 38.0, 25.0], "area": 950.0, "rec": [18, 16, 16, 24, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96], "category_id": 1, "iscrowd": 0, "id": 40, "bezier_pts": [552.0, 193.0, 565.33, 191.33, 577.33, 194.33, 588.0, 202.0, 581.0, 217.0, 571.0, 213.67, 561.0, 210.33, 551.0, 207.0]},

chunchet-ng commented 3 years ago

Hi there,

Sorry for the late reply.

I believe that the format above is taken from the AdelaiDet detection repo, which you have already opened an issue about it.

You can find the definition of original TotalText annotations here.

Thank you.