deeplearning-wisc / stud

source code for CVPR'22 paper "Unknown-Aware Object Detection: Learning What You Don’t Know from Videos in the Wild"
Apache License 2.0
117 stars 14 forks source link

regarding code for inference #1

Closed Jeba-create closed 2 years ago

Jeba-create commented 2 years ago

Thank you so much for such a great work. Could you please share the sample code to do inference on new video samples?

d12306 commented 2 years ago

hi, thanks for your appreciation! @Jeba-create stud does not evaluate on videos, if you want to evaluate on videos, then you can feed single frames in order to stud to get the ood detection performance. that being said, you have to make sure clearly the objects in videos are belonging to id or ood.

for single-frame inference, you can refer to detectron2 doc https://colab.research.google.com/drive/16jcaJoc6bCFAQ96jDe2HwtXj7BMD_-m5 for additional implementation.