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 am working on adding object placement in video which is anchored to scene so if the camera performs any PTZ the position of the object is maintained.
I was wondering if it is possible to tweak the norfair camera_motion module to achieve the above. I'm aware that norfair only does Pan and tilt detection so i'll be working on implementing zoom part separately.
Also wanted general guidance on how i can use the library for my purpose.
The text was updated successfully, but these errors were encountered:
Norfair also allows for zooms since homographies also contemplate zooms. You can use the HomographyTransformationGetter class for the transformation getter of our MotionEstimator for that purpose. There is an example in our camera_motion demo which might be helpful.
Is it possible to use the norfair model on top of some other models than the recommended ones. Specifically i want to use the Segment-anything (SAM) model to get a mask of a surface (not a object) which can also give the bounding box for the same and use the Homography method to place an object there and then track it throughout the video. Will the yolo_dectections_to_norfair_dections be somewhat similar?
I am working on adding object placement in video which is anchored to scene so if the camera performs any PTZ the position of the object is maintained.
I was wondering if it is possible to tweak the norfair camera_motion module to achieve the above. I'm aware that norfair only does Pan and tilt detection so i'll be working on implementing zoom part separately.
Also wanted general guidance on how i can use the library for my purpose.
The text was updated successfully, but these errors were encountered: