Computer Vision & Graphics
Computer-based analysis of images to extract information and classify their contents is becoming increasingly important in all walks of life. For example, by combining the science of 'photogrammetry' (measurement using cameras) with digital camera technology it becomes possible to capture 3D models of people, animals and objects that are metrically accurate and photo-realistic in appearance. Furthermore, it is possible to analyse and animate these models by computer for applications such as virtual actors or sports science.
The Computer Vision and Graphics group, CVG, in the Department of Computing Science, investigates fundamental issues of how to analyse images and also how apply this knowledge within practical applications. Our projects cover all aspects of human body modelling in 3D, including animation and surface skin modelling. This approach opens a wide array of application areas such as; creative media, engineering, medicine, textiles & clothing, military & security, internet & communications, forensic and fine art. A key objective of the work of the group is to combine 3D measurement and modelling techniques with image understanding approaches to construct cognitive robot vision systems that actively search their operating environments using passive digital cameras.
CVG research topics include:
- Medical and veterinary analysis of 3D surface anatomy to assess change following surgical intervention and surgical outcome prediction.
- Object recognition from 2D & 3D information extracted from static images and moving image sequences.
- Whole body scanning and human/animal form modelling, including real-time “immersive” 3D TV.
- Biologically motivated computer vision, including computational models of the mammalian retina and the early visual pathway for efficient and robust image analysis and interpretation.
- Active binocular robot vision systems, able to operate in unstructured and cluttered real-world environments searching and locating visual cues and objects required in autonomous applications such as unmanned vehicle navigation, flexible manufacture, telemedicine and suspicious object inspection.
- New 2D and 3D image compression based on vector quantisation.
- Design of programming languages capable of harnessing parallelism available in vector instruction sets and GPU processors.
- Vector Pascal
- Compilers for the Cell Processor
- Vector Pascal for the Cell
- E# for the Cell
- Oilfield visualisation
- Landmine visualisation using Ground Penetrating Radar
- Physical foundations of computability
Current active research projects in which CVG participates:
- Real-time multi-resolution point-based rendering for display of real-world scanned large datasets.
- Active gaze control and cognitive vision system within a binocular robot sensor head.
- Regularisation of disparity maps based on intensity edge guided anisotropic diffusion.
- Analysis of face range map scans by means of dense vector fields.
- Combined 2D/3D intensity image and range scan analysis of free-form objects exhibiting biological variation.
- Automatic annotation and indexing of image sequences using an artificial retina.
- Development of tools for the procedural generation of architectural structures.
- Compiler optimisation and parallelisation techniques for heterogeneous multi-core processors.
- Code generating techniques that support both processor-level parallelism and data-level parallelism.
- Gloss Perception of Reflectance Models on 3D Surface Textures.
- The Face3D consortium carrying out research into the analysis of three-dimensional facial dysmorphology.
Academic Staff: Dr W Paul Cockshott, Dr J. Paul Siebert.
Research Fellows: Dr John W Patterson.
Research Assistants and Research Students: Mr Gerardo Aragon Camarasa, Mr Sajid Farooq, Mr Youssef Gdura, Mr Paul Graham Keir, Mr Tom Kelly, Mrs Maha Maabar, Miss Susanne B. Oehler, Mr Indradeo Ram, Mr Euan Strachan, Ms Smantha Mullholand, Mr Mathew Dooner.