Part of a computer vision task is to process and analyze the images or video acquired. This includes operations like noise reduction, different forms of filtering and adaptive filtering, multiscale analysis and classification or characterization of object in the images by their shape, size, color, texture and porosity. SINTEF's competence covers algorithm development as well as real-time software and hardware solutions used when implementing the algorithms.

Processing speed is often very important in image processing tasks, especially in online applications. SINTEF has been involved in GPU research for several years, and we always investigate the possibilities to use GPU's (or other dedicated hardware platforms) as accelerating hardware in our projects. A graphics processing unit or GPU (also occasionally called visual processing unit or VPU) is a dedicated graphics rendering device for a personal computer, workstation, or game console. Modern GPUs are very efficient at manipulating and displaying computer graphics, and their highly parallel structure makes them more effective than general-purpose CPUs for a range of complex algorithms.

During the last decade a new branch in image processing techniques has emerged from the mathematical community. These methods can be categorized as variational or PDE (Partial Differential Equations) based image processing and are now starting to be exploited in the industry.

Application examples:

• Computer Vision
• Medical applications
• Process and product inspection
• 3D Shape Measurements and Analysis
• Image enhancement 
• Seismic analysis

The image above shows an unsupervised segmentation of a natural image. Segmentation performed using adaptive statistical models for probabilities, border length regularisation and a  level set formulation. The image features in use are colours and texture.