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 currently using the track method for pose prediction. Will the engine model converted by trtexec also output the ID?I use the track method used by the ultralytics framework to make predictions because this can track people and objects. But how do I get this ID when I convert it to the engine and write the pre- and post-processing by myself?
The text was updated successfully, but these errors were encountered:
@CH-Echo Sorry for late response.
I'm sorry, I don't quite understand what you mean. According to my understanding, tracking algorithms such as DeepSort and ByteTracker assign a unique ID to each target object through a matching strategy, and this ID is not directly obtained from the output of the model. Specifically, detection-based tracking algorithms rely on the detection results output by the object detector and the prediction results of the Kalman filter to achieve tracking. Unless you involve appearance feature matching, then additional deep learning reasoning will be used, otherwise, the tracking algorithm itself does not use additional model inference.
I am currently using the track method for pose prediction. Will the engine model converted by trtexec also output the ID?I use the track method used by the ultralytics framework to make predictions because this can track people and objects. But how do I get this ID when I convert it to the engine and write the pre- and post-processing by myself?
The text was updated successfully, but these errors were encountered: