3DIMPVT 2011: Call for Participation in 3D Urban Data Challenge New Extended Deadline: Feb 28th 2011

Chair: Ioannis Stamos (The City College of New York) http://www.3dimpvt.org/challenge.php

Laser range scanners have now the ability to acquire millions of 3D points of highly detailed and geometrically complex urban sites, opening new avenues of exploration in modeling urban environments. The 3D Urban Data Challenge, in connection with 3DIMPVT 2011 (http://www.3dimpvt.org/), is seeking to attract novel algorithms for processing large point clouds in the city-level.

We will make available two sets of data. The first (NYC) is acquired by Leica ScanStation2. It contains a number of accurate range scans in New York City. The second (LEX) is acquired by a mobile laser range scanner. It includes color + 3D geometry of a downtown strip. Different from the NYC data set, it includes scans from different times of days. This 4D datasets offer more redundant data that may make some processing tasks easier.

We are soliciting solutions in the automated detection and classification of objects of interest or other creative algorithms that use the available datasets. An example problem to tackle is the detection of window features that are prominent in urban scenes, as well as cars on street level, major facades of buildings, balconies, vegetation, street signs, etc. Solutions that explore and detect regularities in urban scenes are of great interest as well. Other creative algorithms for solving problems not mentioned in this document will also be considered.

The contributed solutions will be evaluated by a panel of researchers in the area of range image processing, and the best ideas will be presented at a conference session. A summary  paper will be included in the MVA journal special issue for 3DIMPVT 2011.

Interested party should send an email to the challenge chair (Ioannis Stamos, istamos@hunter.cuny.edu) to request data access.

The deadline for submission of results is February 28th, 2011. Each submission should include a short description about what has been achieved with the data set.