Документ взят из кэша поисковой машины. Адрес оригинального документа : http://graphics.cs.msu.ru/en/research/projects/msr
Дата изменения: Sun Apr 10 00:54:34 2016
Дата индексирования: Sun Apr 10 00:54:34 2016
Кодировка: UTF-8
Computer Vision Collaborative Research in Russia | Graphics and Media Lab

Computer Vision Collaborative Research in Russia

Contact person: Anton S. Konushin (ktosh@graphics.cs.msu.ru)

Overview

Our collaboration with Microsoft Research started from Olga Barinova internship at Microsoft Research Cambridge (MSRC) in 2009. It was continued by a small joint project in sping 2010. In May 2010 we signed a long term contract MRL-2010-050 "Computer Vision Collaborative Research in Russia" for 2010-2014. Our collaboration is not limited to research projects. Microsoft Research sponsored full-length video courses on computer vision, available online now. In 2011 Lomonosov Moscow State University hosted the Microsoft Computer Vision Summer School , which attracted more the 80 students from 30 cities.

Projects

Text detection and recognition in natural images
2010-2013
(1) Automatic detection and understanding the text in natural images, such as photographs of city outdoors or building indoors, is a challenging problem. There is a considerable gap between detecting and understanding text in scanned documents (which is a mature technology) and detecting and understanding text in the natural images. Detection and understanding the text in natural photographs involves localizing the text as well as removing the variation factors, such as varying text orientation, font, color and lighting.
Alpha-flow for video matting
2010-2012
Example description.
Geometric image parsing in man-made environments
2010-2011
We present a new optimization parsing framework for the geometric analysis of a single image coming from a man-made environment. This framework models the scene as a composition of geometric primitives spanning different layers from low level (edges) through mid-level (lines segments, lines and vanishing points) to high level (the zenith and the horizon).
Detection of multiple object instances using Hough transform
2010-2010
In the paper, we develop a new probabilistic framework that is in many ways related to Hough transform, sharing its simplicity and wide applicability. At the same time, the framework bypasses the problem of multiple peaks identification in Hough images, and permits detection of multiple objects without invoking nonmaximum suppression heuristics.

Team

  • Anton S. Konushin, Lomonosov Moscow State University
  • Olga Barinova, Lomonosov Moscow State University
  • Mikhail Sindeev, Lomonosov Moscow State University
  • Victor Lempitsky, SkolkovoTech
  • Pushmeet Kohli, Microsoft Research Cambridge
  • Carsten Rother, Microsoft Research Cambridge

    Acknowledgements

    This project is supported by Microsoft Research programs in Russia.