Open GeekAlexis opened 4 years ago
Here is the activate function: def activate(self, kalman_filter, frame_id): """Start a new tracklet""" self.kalman_filter = kalman_filter self.track_id = self.next_id() self.mean, self.covariance = self.kalman_filter.initiate(self.tlwh_to_xyah(self._tlwh))
self.tracklet_len = 0
self.state = TrackState.Tracked
#self.is_activated = True
self.frame_id = frame_id
self.start_frame = frame_id
as you can see the attribute "is_activated " is not set to True for unconfirmed tracks. In fact, a person must detect at least 2 times to consider as an activated track.
For newly detected tracks, are they marked as unconfirmed or not? In your code, it appears new detections are immediately activated here https://github.com/ifzhang/FairMOT/blob/ca78d3f01d1953c3d72ee8b01f24a5ddec2689c3/src/lib/tracker/multitracker.py#L348 Does it mean there are actually no tracks with the unconfirmed/new track status?