Publications de Stepan Tulyakov

Learning stereo reconstruction with deep neural networks

Stereo reconstruction is a problem of recovering a 3d structure of a scene from a pair of images of the scene, acquired from different viewpoints. It has been investigated for decades and many successful methods were developed. The main drawback of these ...

EPFL2020

Learning an event sequence embedding for event-based deep stereo

François Fleuret, Stepan Tulyakov, Michael Hirsch

Today, a frame-based camera is the sensor of choice for machine vision applications. However, these cameras, originally developed for acquisition of static images rather than for sensing of dynamic uncontrolled visual environments, suffer from high power c ...

IEEE2019

A low-cost, autonomous mobile platform for limnological investigations, supported by high-resolution mesoscale airborne imagery

Ulrich Lemmin, David Andrew Barry, Anton Ivanov, Htet Kyi Wynn, Jérôme Béguin, Ludovic Zulliger, Nawaaz Sidharth Gujja Shaik, Jean-Luc Liardon, Philippe Olivier Paccaud, Beat Marcel Geissmann, Abolfazl Irani Rahaghi, Pascal Klaus, Stepan Tulyakov

Two complementary measurement systems – built upon an autonomous floating craft and a tethered balloon – for lake research and monitoring are presented. The autonomous vehicle was assembled on a catamaran for stability, and is capable of handling a variety ...

2019

Timescales of the Climate Record in the South Polar Ice Cap of Mars

Stepan Tulyakov, Nicolas Thomas

The South Polar Layered Deposits (SPLD) are the largest water ice reservoirs on Mars. Their accumulation is believed to result from climate oscillations that drive the movement of ice and dust on the surface. The High-Resolution Imaging Science Experiment ...

AMER GEOPHYSICAL UNION2019

Practical Deep Stereo (PDS): Toward applications-friendly deep stereo matching

François Fleuret, Anton Ivanov, Stepan Tulyakov

End-to-end deep-learning networks recently demonstrated extremely good performance for stereo matching. However, existing networks are difficult to use for practical applications since (1) they are memory-hungry and unable to process even modest-size image ...

NEURAL INFORMATION PROCESSING SYSTEMS (NIPS)2018

Geometric calibration of Colour and Stereo Surface Imaging System of ESA's Trace Gas Orbiter

François Fleuret, Anton Ivanov, Thomas Weigel, Stepan Tulyakov, Nicolas Thomas

There are many geometric calibration methods for "standard" cameras. These methods, however, cannot be used for the calibration of telescopes with large focal lengths and complex off-axis optics. Moreover, specialized calibration methods for the telescopes ...

Elsevier2018