Ústav automobilovej mechatroniky srdečne pozýva všetkých záujemcov na prednášku "Self-driving Cars: Deep Learning, Computer Vision and Sensor Fusion". Prednášať bude Peter Škvarenina z JetBrains.
Prednáška sa bude konať v stredu 28. 6. 2017 13:00 - 15:00 v miestnosti CD150 na FEI STU, Ilkovičova 3, Bratislava. Prednes bude v angličtine.
Abstract: The goal of this presentation is to provide a guide to accessible yet very powerful examples on how to program own self-driving cars. The presentation is divided into four sections. First section covers an end-to-end approach to self-driving cars using Deep Learning (Convolutional Neural Networks), based on NVidia's research as utilized by Roborace race cars, Tesla or Audi. Second section covers the topic of how to program an advanced lane detection using computer vision and OpenCV. Third section concerns itself with vehicle detection using computer vision based on Histogram of Oriented Gradients and linear SVM classifier. Fourth and last section covers sensor fusion of radar and LiDAR data for tracking objects such as pedestrian using Extended Kalman Filters.
Author: Peter Skvarenina graduated from Faculty of Management and Computer Science of University of Zilina and currently continues his studies at Georgia Institute of Technology in Atlanta in Machine Learning and Robotics, while living in Frankfurt am Main in Germany. In the past Peter worked for some of the most accomplished technological companies on various topics, such as on transacted distributed systems with high availability at SUN Microsystems Inc., at NOKIA on automotive industry's 3D navigation standard, planetary-wide 3D visualization and Big Data processing, and currently at JetBrains, where he prepared many algorithms (mainly geometric and computer vision) for Animatron, an online animation editor.