Purpose

GeoCast is a data format that binds together camera information and multi-layer image sequences, such as RGBZ or RGBA representations. We use OpenEXR for image storage, and represent the camera position, rotation, and lens parameters as text files.

This site hosts the documentation and helper software for GeoCast, a camera parameter standard for the exchange of extrinsic and intrinsic camera parameters (even dynamic) across applications.

It is especially useful in applications where computer vision and computer graphics need to work together to create new data from 2D imagery (aka Visual Computing).

Background

GeoCast : A dynamic camera parameter specification. Gernot Ziegler, Lukas Heidenreich, Marcus Magnor and Hans-Peter Seidel GeoCast: Unifying Depth Video with Camera Meta-Data In: 2nd Workshop on Immersive Communication and Broadcast Systems, Berlin, Germany, Heinrich-Hertz-Institute, Berlin, 2005, 0-4 Abstract: We present a storage format for placing multiple, dynamic 2.5D point sample streams (e.g. RGBZ video) or color video projections into the context of a common world space. The approach utilizes the concept of projective geometry to let virtual projectors "cast" the data from where it was originally recorded by the involved cameras. We exemplify the data format's versatility by demonstrating how several moving cameras reproduce 3D geometry exported from modeling software, and outline the extension to real-world acquisition. We also explain how this format can act as common interchange format for the camera parameters of lightfield/multi-view video footage.

Presentations

Documentation

GeoCast 3ds Max Camera Exporter 2.0

GeoCast Maya Camera Exporter 0.1

GeoCast Blender Camera Exporter 0.15

Downloads

You can download the exporter scripts for GeoCast here. For the documentation, this refer to the Wiki.

GeoCast Blender Camera Exporter 0.15
GeoCast 3ds Max Camera Exporter 2.01
GeoCast Maya Camera Exporter 0.1
Matlab to GeoCast Camera Converter

Note: For using the Matlab to GeoCast Camera Converter, you'll need the Camera Calibration Toolbox for Matlab.

Demo material

• There is a short movie available that demonstrates the use of libgeocast in an application that uses the data to regenerate the modelled volume by only using rendered images and the geocast files. You can see that the geometry is reconstructed quite accurate.
  
  
  Short demo movie(MPEG1, 6 MB).
  
  
• This GeoCast/OpenEXR sequence features the synthetic model of a Volkswagen Sedan seen through an orthographic camera which pans around the frontside of the car. geocast/ contains the GeoCast files which describe the camera motion and project img/ contains OpenEXR images of the video sequence, including a Z channel. The .geoscene file describes the actual composition of GeoCast and OpenEXR files into a video sequence. The 3D model and camera path is available in the accompanying Blender file. The GeoCast generation script for Blender's Python engine is available on request.

Related sites

• The EU research project 3D Television
• Video Based Rendering Link Page
• Lukas Heidenreich's GeoMerge page
• GeoCast Wiki

Discussion

Pleae join the mailing list geocast-devel.

NEWS:

• 20070410: created wiki.
• 20060311: added movies, presentations, mailing list.
• This page has been created.