Open harshal-14 opened 2 months ago
The motion can be in any direction.
Ttc is generalized to all 3 directions in equation 1. Then Theorem 3.1 and 3.2 provide an equivalent constraint for each of the 3 directions.
How can I integrate an ML detection model(YOLOv9 or RT-DETR) for detection of objects rather than just patches?
Supposedly the camera motion is only forward direction and not sideways/backward, the algorithm doesn't estimate the ttc. Is there a workaround this or is this a limitation of the proposed method?