Missing information for LVIS dataset

facebookresearch / detectron2

Detectron2 is a platform for object detection, segmentation and other visual recognition tasks.

https://detectron2.readthedocs.io/en/latest/

Apache License 2.0

30.55k stars 7.49k forks source link

Missing information for LVIS dataset #670

Closed ggalan87 closed 4 years ago

ggalan87 commented 4 years ago

I noticed a problem in lvis api wrapper. From lvis spec the expected fields for annotations are 'id', 'area', 'bbox', 'bbox_mode', 'category_id', 'segmentation', however both 'id' and 'area' are discarded in the code below

https://github.com/facebookresearch/detectron2/blob/4f68200db39ce43d58a97c42e6fc82075a843cea/detectron2/data/datasets/lvis.py#L125-L142

ppwwyyxx commented 4 years ago

they are discarded because they are not useful to us