A Computational Approach to Edge Detection
IEEE Transactions on Pattern Analysis and Machine Intelligence
Recovery of Parametric Models from Range Images: The Case for Superquadrics with Global Deformations
IEEE Transactions on Pattern Analysis and Machine Intelligence
Dynamic 3D Models with Local and Global Deformations: Deformable Superquadrics
IEEE Transactions on Pattern Analysis and Machine Intelligence
A unified distance transform algorithm and architecture
Machine Vision and Applications
Multiresolution stochastic hybrid shape models with fractal priors
ACM Transactions on Graphics (TOG) - Special issue on interactive sculpting
New feature points based on geometric invariants for 3D image registration
International Journal of Computer Vision
Genetic Algorithms in Search, Optimization and Machine Learning
Genetic Algorithms in Search, Optimization and Machine Learning
Superquadrics and Free-Form Deformations: A Global Model to Fit and Track 3D Medical Data
CVRMed '95 Proceedings of the First International Conference on Computer Vision, Virtual Reality and Robotics in Medicine
Localization of 3D Anatomical Point Landmarks in 3D Tomographic Images Using Deformable Models
MICCAI '00 Proceedings of the Third International Conference on Medical Image Computing and Computer-Assisted Intervention
| Hi-index | 0.00 |
Existing approaches to extracting 3D point landmarks based on deformable models require a good model initialization to avoid local suboptima during model fitting. To overcome this drawback, we propose a generally applicable novel hybrid optimization algorithm combining the advantages of both conjugate gradient (cg-)optimization (known for its time efficiency) and genetic algorithms (exhibiting robustness against local suboptima). We apply our algorithm to 3DMR and CTimages depicting tip-like and saddle-like anatomical structures such as the horns of the lateral ventricles in the human brain or the zygomatic bones as part of the skull. Experimental results demonstrate that the robustness of model fitting is significantly improved using hybrid optimization compared to a purely local cg-method. Moreover, we compare an edge strength- to an edge distance-based fitting measure.