Research

My research interests are founded in image analysis and in particular medical image analysis and can be categorized under the three main headings below.

Shape Model Building

My work on shape models has included medial models, traditional point distribution models, and hybrids thereof. The main goal is to automatically generate statistical shape models from annotated training data - in 3D in the absense of annotated anatomical landmarks. The driving applications have lately been modeling of the knee joint and prior to that to produce a prostate shape model.

Machine Learning

I dont really consider myself a "machine learning researcher". However, as pattern recognition methodology becomes more and more integrated in every line of research, most of my research has some aspect of machine learning.

The focus has been efficient voxel classification, feature selection/combination, and determination of the proper set of training cases.

Musculoskeletal Imaging Markers

In the musculoskeletal projects at BiomedIQ, we collaborate with Nordic Bioscience (Herlev, Denmark), the Center for Clinical and Basic Research (CCBR, Ballerup, Denmark), and Synarc (San Francisco, USA). The focus is to develop quantitative marker of disease progression based on automated image analysis.