Could you provide details of the test using the object detector of DRG?

[ddwhzh]

I took the liberty of asking the details of the object detector of DRG used by Table 5 GGNet‡.
Is it just the results of Faster Rcnn or the results with the human and object detection branches?
Any code or anno files would be greatly appreciated.

[SherlockHolmes221]

We use this object detection results test_hico_finetune_drg.json ( added in the annotation links just now).
The file is get from DRG.
Any more questions?

[ddwhzh]

thank you