The most used pre-trained networks for computer vision (e.g. ResNet50) are trained on 3 channels (RGB). At the same time, many cameras used in robotics return RGB-D outputs, that is including depth information.

Fine tuning 3-channels pre-trained network on RGB-D outputs force us to give up one of the channels. Instead, is there a publicly available vision pre-trained network trained on RGB-D (4) channels?

  • 1
    $\begingroup$ Asking for pre-trained models here is off-topic. We focus on the theoretical aspects of AI. This question is probably more suitable for Data Science SE. $\endgroup$
    – nbro
    Dec 21 '20 at 17:36