Unsupervised Segmentation of 3D Images (GR/K 45487)

1. Development of an unsupervised segmentation algorithm, applicable to 2D and 3D data based on a hierarchical region analysis using region and edge information. [2]
2. Development of an unsupervised segmentation algorithm, applicable to 2D and 3D data, using a region growing approach based on a robust local clustering technique. [4, 6, 9]
3. Development of a novel feature extraction technique for 3D seismic images based on local instantaneous frequency and bandwidth. [4, 5]
4. Development of a supervised image segmentation algorithm using a multi-scale surface model based on closed B-splines. [3, 7]
5. Development of a novel data structure for representing 3D labelled regions and a quantitative analysis of algorithmic performance in comparison with the octree for 3D region operations. [1]
6. Quantitative evaluation of algorithms on 3D seismic and 2D range image data. [6, 8]

For more information please contact: