Submitted by Nikolai Smolyanskiy.

Submission data

Full name	NVStereoNet: An Efficient Semi-Supervised Deep Neural Network Approach
Description	We propose a novel semi-supervised learning approach to training a deep stereo neural network, along with a novel architecture containing a machine-learned argmax layer and a custom, publicly available runtime that enables a smaller version of our stereo DNN to run on an embedded GPU.
Publication title	On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach
Publication authors	Nikolai Smolyanskiy, Alexey Kamenev, Stan Birchfield
Publication venue	Arxiv
Publication URL	https://arxiv.org/abs/1803.09719
Programming language(s)	TensorFlow
Hardware	Titan XP
Submission creation date	1 Jun, 2018
Last edited	1 Jun, 2018

High-res multi-view results

Low-res many-view results

Low-res two-view results

SLAM results