ForK - Building bridges between state-of-the-art research and creative media

Hochschule Bonn-Rhein-Sieg
Transition from recent research implementations to mobile web technologies.
Departments and Instituts: 
Computer Science, Institute of Visual Computing
From: 
01.09.2016
to
31.08.2019
Project website: 
IVC project page

Project description

Hochschule Bonn-Rhein-Sieg

The latest developments in research are often not directly applicable for the media and creative industries, since the prototypical implementations are often operated using special software and hardware environments that are not designed for common platforms. In contrast, usage of web technology is basically a standard in the media and creative industries.

For this reason, the ForK project aims to develop a process which makes the transition from current research implementations to mobile web technologies as simple as possible. A (partially) automated conversion chain is modeled and examined, in which a machine-oriented programming language such as C++ is transferred to the browser-based programming language Javascript using the transpiler Emscripten. In particular, aspects of the two programming languages, which have an influence on a successful conversion, will be examined.

The Institute of Visual Computing is developing a research prototype from the area of ​​Augmented Reality which will be used to continuously test and optimize the transition process defined by the project partners.

Project manager at H-BRS

Prof. Dr. André Hinkenjann

Gründungsdirektor des Institut für Visual Computing
Forschungsprofessur für Computergrafik und Interaktive Systeme
Hochschule Bonn-Rhein-Sieg
Email: 
andre.hinkenjann [at] h-brs.de

Research associates

Results

The scalable research prototype which is being developed at the H-BRS can handle various aspects from the field of augmented reality.

This involves the following components:

  • 3D point cloud reconstruction of a captured scene using structure-from-motion.
  • A tracking system which uses reference information to find the position and orientation of a camera.
Hochschule Bonn-Rhein-Sieg
With structure-from-motion a 3D point cloud is generated from several input images.
Hochschule Bonn-Rhein-Sieg
The external camera parameters of a known scene can be determined, so that the corresponding 3D model can be superimposed on the image.

Large parts of the implemented library could be transformed to JavaScript. We used the resulting JavaScript library to create several web demos using the transpiled functionality.

 

Hochschule Bonn-Rhein-Sieg
One of the resulting web demos. The JavaScript library which has been transpiled from C++ code is used to demonstrate image distortion and undistortion using the required parameters.

The results in form of source code, examples and documentation can be found here:

Sponsors

Co-operation partners