Present Projects

Clinical Applications and Theory of Image Analysis

CATIA is a collaboration between the research institute Nordic Bioscience and the Image Group at University of Copenhagen. The aim of the collaboration is to make the medical know-how and practical experience of Nordic Bioscience available for the theoretical research that goes on at University of Copenhagen. Nordic Bioscience contributes with access to their ressources as well as research grants for researchers at University of Copenhagen.
Duration: 2004-2008

Contact: Mads Nielsen

Computer-Aided Assessment of COPD from CT Images

The project is a joint initative of the Image Group at University of Copenhagen and Gentofte University Hospital in Copenhagen. Chronic obstructive pulmonary disease (COPD) is a major cause of morbidity and mortality worldwide. The mechanisms involved in COPD are not yet well understood and a cure is currently not available. The research project aims to a) develop and evaluate image processing techniques to enable reproducible, localized measurements of lesion morphometry and progression, b) study disease development and progression in a large group of smokers currently enrolled in the Danish lung cancer screening trial, and c) study disease characteristics in patients with rapid decline compared to non-rapid decline of lung function. This would aid in improving our understanding of pathogenesis of COPD, enable a better risk stratification of the disease, possibly in an earlier, pre-clinical stage, and facilitate monitoring of therapy.
Duration: 2006-2010

Contact: Marleen De Bruijne

Natural Image Sequence Analysis

NISA is a research collaboration between the Image Group at University of Copenhagen and University of Southern Denmark. The partners will study image sequences as they appear from natural scenes by statistical and geometrical means. Statistical analysis of still images has shown fruitful results both for solving practical applications, but also more importantly in this context, to gain insight into the nature of images. This line of research to image sequences will be furthered in the project along several lines answering questions like: What are the natural building blocks of image sequences? What are the natural motion patterns? What is the descriptive power of local spatio-temporal features? To which extent may they be used for motion analysis?
Duration: 2006-2008

Contact: Kim Steenstrup Pedersen

Quantitative Shape Modeling in Biomedical Imaging

The research consortium QSMBI is a joint effort between industry and academia with a three years grant from the Danish Technical Research Council. Shape modeling is an essential ingredient in image analysis. Especially in the biomedical area, where shapes are not manmade, they are complex and often exhibit strange and large variability. Statistical shape models are used both for inference of objects shape in images and for classification as diagnosis. The goal of the consortium is to develop and evaluate the descriptive power of new shape modeling techniques by solving demanding and important real world applications in the biomedical area.
Duration: 2005-2008

Contact: Mads Nielsen

See also: QSMBI project homepage
Computational and Cognitive Vision Systems: A Training European Network

Visiontrain is a Marie Curie research training network supported by the EU IST 6th framework program. Visiontrain addresses the problem of understanding vision from both computational and cognitive points of view. The intension is to reduce the gap that exists today between biological vision (which is by large not yet understood) and computer vision (which is biologically inspired and whose flexibility, robustness, and autonomy remain to be demonstrated). There are a total of 11 partners.
Duration: 2005-2009

Contact: Mads Nielsen

See also: VISIONTRAIN project homepage

Ended Projects

Clinical Applications and Theory of Image Analysis

Among the highest ranking death causes for women in the western world are breast cancer and cardiovascular diseases. Furthermore, osteoporosis is present in large number of women after the menopause. The aim of this project is to develop more objective and quantitative measures of related body status based on x-rays. This may lead to better diagnosis, risk assessment, and treatment. Furthermore, due to the extremely large amount of data in this project it may lead to image analysis methodologies that can be transferred to other image analysis problems.
Duration: 2002-2006

Contact: Mads Nielsen

Deep Structure, Singularities, and Computer Vision

The project is sponsored by the IST Programme of the European Union. The goal of the project is to contribute to the scientific and technological development in computer vision using deep structure singularities as data structures and efficient algorithms for operations on these. By the end of the project, shape and image descriptors efficiency for solving computer vision task have been analysed and tested on image matching and database indexing, especially in the context of medical image archives. As a result of the project, efficient image database searches will have been developed in the context of MR scans.
Duration: 2002-2005

Contact: Mads Nielsen

Natural Shape

Objects in nature exhibit a complexity in shape and appearance making artificial recognition, representation, and communication extremely difficult. In this center computer scientists, mathematicians, psychologists, and medical doctors work together on aspects of modelling shape and appearance of natural objects and phenomena. One of the major challenges is to represent complicated structures in a finite memory in a generic fashion.
Duration: 2002-2005

Contact: Peter Johansen

See also: Natural Shape project homepage