Intelligent clustering with instance-level constraints
Intelligent clustering with instance-level constraints
White matter tract clustering and correspondence in populations
MICCAI'05 Proceedings of the 8th international conference on Medical Image Computing and Computer-Assisted Intervention - Volume Part I
Automated atlas-based clustering of white matter fiber tracts from DTMRI
MICCAI'05 Proceedings of the 8th international conference on Medical Image Computing and Computer-Assisted Intervention - Volume Part I
Agglomerative hierarchical clustering with constraints: theoretical and empirical results
PKDD'05 Proceedings of the 9th European conference on Principles and Practice of Knowledge Discovery in Databases
| Hi-index | 0.00 |
In this work we introduce the novel concept of applying constraints into the fiber segmentation problem within a clustering based framework. The segmentation process is guided in an interactive manner. It allows the definition of relationships between individual and sets of fibers. These relationships are realized as pairwise linkage constraints to perform a constrained clustering. Furthermore, they can be refined iteratively, making the process of segmenting tracts quicker and more intuitive. The current implementation is based on a constrained threshold based clustering algorithm using the mean closest point distance as measure to estimate the similarity between fibers. The feasibility and the advantage of constrained clustering are demonstrated via segmentation of a set of specific tracts such as the cortico-spinal tracts and corpus collosum.