no code implementations • 18 Dec 2017 • Martin Velas, Michal Spanel, Michal Hradis, Adam Herout
Our networks show significantly better precision in the estimation of translational motion parameters comparing with state of the art method LOAM, while achieving real-time performance.
Robotics
no code implementations • 7 Sep 2017 • Martin Velas, Michal Spanel, Michal Hradis, Adam Herout
This paper presents a novel method for ground segmentation in Velodyne point clouds.
Robotics
no code implementations • 12 Jul 2016 • Martin Cadik, Jan Vasicek, Michal Hradis, Filip Radenovic, Ondrej Chum
This work addresses the problem of camera elevation estimation from a single photograph in an outdoor environment.
1 code implementation • 2 May 2016 • Pavel Svoboda, Michal Hradis, David Barina, Pavel Zemcik
This paper shows that it is possible to train large and deep convolutional neural networks (CNN) for JPEG compression artifacts reduction, and that such networks can provide significantly better reconstruction quality compared to previously used smaller networks as well as to any other state-of-the-art methods.
no code implementations • 25 Feb 2016 • Pavel Svoboda, Michal Hradis, Lukas Marsik, Pavel Zemcik
In this work we explore the previously proposed approach of direct blind deconvolution and denoising with convolutional neural networks in a situation where the blur kernels are partially constrained.