Detection of malignant melanoma based on lesion images

Computer-aided diagnostics (CAD) systems are already widely used in practical clinical situations for tasks such as the detection of breast and lung cancer and melanoma.

Background

Computer-aided diagnostics (CAD) systems are already widely used in practical clinical situations for tasks such as the detection of breast and lung cancer and melanoma.  Most of these systems are detection systems designed to aid physicians when looking for abnormalities in diagnostic data.

Goal

The main goal is to contribute to this research field by developing analysis methods that are suitable for fully automatic and scalable differential systems involving statistical inference for the purpose of diagnostic decision making.  Since the goal of the project is to deliver the system to a wide population, the use of mobile devices will be investigated.

Problems

The challenges lie within automatic image segmentation that is robust and fast enough and in pattern recognition methods that can handle anatomical structures in images.  Analysis methods also need to be tailored for images acquired by low-cost devices such as pocket-dermatoscopes and mobile phone digital cameras.

Methods

Existing segmentation techniques, Scale-space methods, and Gaussian Markov Random Fields will be applied.  New robust methods suitable for this particular problem will be designed.  Clinical trials to measure performance of the system.

Project members from partners

Prof. Fred Godtliebsen (UiT), PhD Kristian Hindberg (UiT)

Researchers

1 Post.doc researcher and 1 PhD student

International collaborators

Prof. Lasse Holmström, Department of Mathematics and Statistics, University of Oulo, Finland, Dr. Jörg Polzehl, Weierstrass Institut für Angewandte Mathematik und Statistikk, Berlin, Germany Prof. Probal Chaudhuri, Indian Statistical Institute, Calcutta, India Prof. James Stephen Marron, University of North Carolina at Chapel Hill, Chapel Hill, USA.

Contribution to health care

Diagnostic equipments that can be used by dermatologists in order to exploit resources in a more efficient way.

Contribution to new industry

Software that accompanies hardware products and that creates basis for new services.

Project participant

Kajsa Møllersen, e-mail: email hidden; JavaScript is required, tel.: +47 977 83 940.