Kitti dataset preprocess without images

[DJNing]

Since this is a Lidar-only object detection model, could we preprocess the data without the images? I think if we don't use the evaluation metric of 'bbox', we can ignore the infos from images, right?

Another question is, why do we only count the points inside the gt in the FOV of the image?

[tianweiy]

Since this is a Lidar-only object detection model, could we preprocess the data without the images? I think if we don't use the evaluation metric of 'bbox', we can ignore the infos from images, right?

I think so ?

Another question is, why do we only count the points inside the gt in the FOV of the image?
only these boxes are annotated

[DJNing]

Thanks for the reply, I think I've made it work now. One more question, how to change the number of target classes for the network? Can I just change the CLASS_NAMES in the network YAML directly?

[tianweiy]

yeah, i think this should work

[francescofact]

Thanks for the reply, I think I've made it work now. One more question, how to change the number of target classes for the network? Can I just change the CLASS_NAMES in the network YAML directly?

How did you managed to preprocess the kitti dataset without the calibration images? I also have a lidar only dataset