Chandramouli Rajagopalan

Can I work on this issue ?

I have not trained it on KITTI dataset, but I have applied the same procedure on a different private dataset where I got good results. I will train it and...

This class https://github.com/chamorajg/pytorch_depth_and_motion_planning/blob/master/depth_prediction_net.py#L48 should give you the depth map as an output.

I haven't used it for Kitti or Cityscape, but all you need to do is make a sequence pair of images and share the path to those files.

I will add a license file to this. You can experiment with this project, I don't have any issue with it.

I have used the RGBD image from the model and the intrinsic parameters and then tried to reconstruct using [Open3D](http://www.open3d.org/).

The _mask function is used as thresholding to filter out the regions where L2 norm is greater than the average of all the regions.

Licensed under MIT license.