I am an experienced driver, instructor and also a microelectronics engineer working with an autonomous system on topic of pathing & control. I see in your videos that you show the view of your color camera input and I see here that you are showing pathing models trained directly on color camera information. Do you do some hand-written processing of the camera image for training and whatever path detection you do? Models in my team work on images processed by feature extraction algorithms, which allow for the path identification models to be much smaller, faster and precise.
![[CVPR2024 presentation] OmniLocalRF: Omnidirectional Local Radiance Fields from Dynamic Videos](/img/n.gif)
Amazing watch! Thank you for great presentation
Great Job!!
I am an experienced driver, instructor and also a microelectronics engineer working with an autonomous system on topic of pathing & control.
I see in your videos that you show the view of your color camera input and I see here that you are showing pathing models trained directly on color camera information.
Do you do some hand-written processing of the camera image for training and whatever path detection you do?
Models in my team work on images processed by feature extraction algorithms, which allow for the path identification models to be much smaller, faster and precise.