How to detect bounding boxes at the back of the vehicle?

[vardeep-sandhu]

This is a dataset/general query.

I have trained an object detector with "centerpoint_rnn.cfg" config and the model has trained and is working well. But I noticed just now that the KITTI dataset has gt bbox information just for objects which are in the FOV of the cameras. So the cars at the back of the vehicle are not detected. This has caused problems for me since I am working on Moving Object Segmentation and I want to detect vehicles that are at the back as well. Any suggestion on how that can be achieved.

In the image, it can be seen that the cars in front of the vehicle bearing laser scanner are detected but there is a car just behind the vehicle which remains undetected

[tianweiy]

Hi, I am a little confused. So do you have any other dataset to train / test on that also have cars not in fov? Could you use their label then ? I know semantic kitti has label for this?

If you have to use model trained on KITTI, maybe you can change the point cloud range to make sure the detector see the car behind it

[vardeep-sandhu]

Yes, changing the point cloud range worked. I didn't know how it was being used. So the detector was not able to see the cars behind. But now it is working.

Thanks for the response tho

[maayanYa]

Hi ,
@here-to-learn0 I'm having the same issue but somehow changing the point cloud range didn't effect .
The detector doesn't recognise the objects in the right part of the point cloud and even when I change the point cloud range and make sure the point coordinates are inside the point cloud range - objects there are still not detected at all.

And I also noticed that not every number there is accepted, some cause bugs , didn't understand yet what causes it .

Did you change anything else other than the point cloud range ?

Thank you !