You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've been exploring the tutorials available for predicting masks using Mask R-CNN. As Mask R-CNN can also be extended to include pose estimation, I'm interested in whether there are similar resources or scripts available.
Specifically, I'm looking for a script similar to create_onnx.py that not only handles mask prediction but also supports pose estimation alongside the detection of classes and bounding boxes. This would greatly help in applications that require a comprehensive understanding of both object identity and pose.
Could you provide guidance or resources for implementing Mask R-CNN with integrated pose estimation features? Any help or pointers to scripts that handle these capabilities would be much appreciated.
The text was updated successfully, but these errors were encountered:
I understand, but it looks like they'll need to update the 'create_onnx.py' script, especially the roi_head function, to include pose estimation. We'll have to wait and see if they decide to implement these changes.
Hello!
I've been exploring the tutorials available for predicting masks using Mask R-CNN. As Mask R-CNN can also be extended to include pose estimation, I'm interested in whether there are similar resources or scripts available.
Specifically, I'm looking for a script similar to create_onnx.py that not only handles mask prediction but also supports pose estimation alongside the detection of classes and bounding boxes. This would greatly help in applications that require a comprehensive understanding of both object identity and pose.
Could you provide guidance or resources for implementing Mask R-CNN with integrated pose estimation features? Any help or pointers to scripts that handle these capabilities would be much appreciated.
The text was updated successfully, but these errors were encountered: