Image processing for integrated and immersive visualisation

Project Details

Description

The field of visualization studies transformations that aim to augment the understanding and the information contents of multidimensional data for a typical end-user. An important source of such data are multidimensional digital images. These images can be dense (e.g. 3D medical images) or sparse (e.g. 3D CAD scene descriptions) or a combination of both. Miscellaneous image data transformations at different processing levels do exist for visualization. This project identifies four research areas related to image processing on multidimensional data for integrated and immersive visualization purposes. The organization of the proposal into four workpackages reflects these four generic research areas:

Image compression.
Integrated and immersive visualisation applications encounter their limitations in data communication between underlying hardware architectures. Powerful and flexible compression techniques are required to reduce the consumed communication bandwidth to a minimum. In this context, bandwidth refers as well to the memory size as to the transmission capacity of internal data buses and of external communication networks. This proposal addresses a number of techniques to improve compression efficiency such as progressive volumetric compression, texture coding for dynamic scenes and model-based compression for video sequences.

Analysis, reconstruction and interpretation of image data.
New and enhanced analysis techniques are developed to transform a multitude of image data formats (video, (multi-modal) 3D volumes, mesh and scattered data) in higher level descriptions. Such transformations enable e.g. integrated visualization by image fusion. Problems will be solved in the area of motion based segmentation of objects in video scenes, wavelet based visualisation of multimodal texture images, fitting of surfaces to scattered data, and 3D mesh simplification.

Virtual and augmented environments: integration, (haptic) interaction and co-operation evaluation.
Visualization methods are migrating more and more towards integrated methods in which the visual presentation not only comprises purely synthesized images, but also includes synthesized images correctly overlayed with real-life images. Moreover, the interaction metaphor in these 3D virtual environments is clearly outgrowing the purely visual representation by including auditive and haptic feedback in the interaction as well. This proposal addresses overlay techniques (image-graphics), new forms of interaction with image data, immersive co-operation between actors in virtual 3D scenes and integrated imaging and presentation for image-guided manipulation. It is an overall objective to perform research for suitable evaluation methods of techniques in this context.

Implementation and realisation.
Given the ever-increasing (implementation) complexity of integrated and immersive visualisation systems, it becomes essential that application and system designers obtain better support to make accurate and rapid implementation decisions at a high level. In this proposal research concerning automatic generation of system-level cost estimators and the guaranteeing of a pre-defined Quality of Service will be performed, driven by the developments originating from the first two research areas.
AcronymIWT68
StatusFinished
Effective start/end date1/07/9930/06/01

Flemish discipline codes

  • Electrical and electronic engineering
  • Mathematical sciences

Keywords

  • visualization
  • Visualisation
  • Image compression
  • Image reconstruction
  • Image Analysis