Multiple ROI heads for predicting multiple attributes as in bottom-up attention

facebookresearch / Detectron

FAIR's research platform for object detection research, implementing popular algorithms like Mask R-CNN and RetinaNet.

Apache License 2.0

26.25k stars 5.45k forks source link

Multiple ROI heads for predicting multiple attributes as in bottom-up attention #978

Closed ggaemo closed 4 years ago

ggaemo commented 4 years ago

I'm trying to use detectron for predicting not only the bounding box and object label but also for other attributes.

For instance, I'm trying to predict the shape, angle and color of the object as well. How can I modify the yaml file or model code? Are there any reference I can refer to? Is it possible in the first place?

Thank you very much.