Untitled Document
Not a member yet? Register for full benefits!

 VideoTrace: Rapid interactive scene modelling from video

This story is from the category Graphics
Printer Friendly Version
Email to a Friend (currently Down)



Date posted: 29/09/2007

A new modelling system, VideoTrace; the result of a collaboration between The Australian Centre for Visual Technologies at the University of Adelaide, and The Oxford Brookes Computer Vision Group, is capable of taking the output of any handheld digital camcorder, and turn it into a 3D model.

The user takes a film of the object to be recreated from every possible angle ? just walking round it, camera in hand, pointing at the object as they move.

This film file is then uploaded into VideoTrace. They then trace tracing the shape of the object to be modelled over one or more frames of the video. The video can be played forwards or backwards, and the software is intelligent enough to work out which parts of the object to anchor to, and moves the sketched out areas with the car as position changes. This allows the sketch to become truly 3D as new parts are added from different angles.

By interpreting the sketch drawn by the user in light of 3D information obtained from computer vision techniques, the software can work out if a surface is horizontal or vertical, adjusting the plains of the sketch to suit, and working with the user, as more points and curves are added over time, to complete the 3D model.

VideoTrace provides an intuitive and powerful means of modelling shape from video, and executes quickly enough to be used interactively.

The combination of automated and manual reconstruction allows VideoTrace to model parts of the scene not visible, and to succeed in cases where purely automated approaches would fail.

See the full Story via external site: www.acvt.com.au

Most recent stories in this category (Graphics):

07/02/2017: Complex 3D data on all devices

06/05/2014: U-M paleontologists unveil online showcase of 3-D fossil remains

17/03/2014: 3D X-ray Film: Rapid Movements in Real Time

07/02/2014: Modelling the Duynamics of the Skin

20/01/2014: CCNY Team Models Sudden Thickening of Complex Fluids

12/11/2013: Visualizing the past: Nondestructive imaging of ancient fossils

14/08/2013: Shadows and light: Dartmouth researchers develop new software to detect forged photos

05/08/2013: Seeing depth through a single lens