The KITTI Vision Benchmark Suite

Depth estimation is an important capability for
autonomous vehicles to understand and reconstruct
3D environments as well as avoid obstacles during
the execution. Accurate depth sensors such as
LiDARs are often heavy, expensive and can only
provide sparse depth while lighter depth sensors
such as stereo cameras are noiser in comparison.
We propose an end-to-end learning algorithm that
is capable of using sparse, noisy input depth for
refinement and depth completion. Our model also
produces the camera pose as a byproduct, making
it a great solution for autonomous systems. We
evaluate our approach on both indoor and outdoor
datasets. Empirical results show that our method
performs well on the
KITTI~\cite{kitti_geiger2012we} dataset when
compared to other competing methods, while having
superior performance in dealing with sparse,
noisy input depth on the TUM~\cite{sturm12iros}
dataset.

Our algorithm is implemented with Pytorch. We use
ADAM optimizer with a momentum of 0.9 and weight
decay of 0.0003. The ADAM parameters are pre-
selected as $\alpha = 0.0001, \beta_1 = 0.9$ and
$\beta_2 = 0.999$. Two Tesla V100 with RAM of
32GB are used for training with the batch size of
8 and 15 epochs take around 12 hours.

@article{Zhang2019DFineNetEE,
title={DFineNet: Ego-Motion Estimation and
Depth Refinement from Sparse, Noisy Depth Input
with RGB Guidance},
author={Yilun Zhang and Ty Nguyen and Ian D.
Miller and Shreyas S. Shivakumar and Steven W.
Chen and Camillo J. Taylor and Vijay Kumar},
journal={CoRR},
year={2019},
volume={abs/1903.06397}
}