Virtual craniofacial reconstruction using computer vision, graph theory and geometric constraints

Authors:
Ananda S. Chowdhury;Suchendra M. Bhandarkar;Robert W. Robinson;Jack C. Yu
Affiliations:
Department of Computer Science, The University of Georgia, Athens, GA 30602, USA;Department of Computer Science, The University of Georgia, Athens, GA 30602, USA;Department of Computer Science, The University of Georgia, Athens, GA 30602, USA;Department of Plastic Surgery, The Medical College of Georgia, Augusta, GA 30912, USA
Venue:
Pattern Recognition Letters
Year:
2009

Citing 7
Cited 0

A survey of thresholding techniques

Computer Vision, Graphics, and Image Processing
3-D Object Recognition Using Bipartite Matching Embedded in Discrete Relaxation

IEEE Transactions on Pattern Analysis and Machine Intelligence
A Method for Registration of 3-D Shapes

IEEE Transactions on Pattern Analysis and Machine Intelligence - Special issue on interpretation of 3-D scenes—part II
Algorithms on Trees and Graphs

Algorithms on Trees and Graphs
Assessing Craniofacial Surgical Simulation

IEEE Computer Graphics and Applications
Comparing Images Using the Hausdorff Distance

IEEE Transactions on Pattern Analysis and Machine Intelligence
Rigid Point-Surface Registration Using an EM Variant of ICP for Computer Guided Oral Implantology

MICCAI '01 Proceedings of the 4th International Conference on Medical Image Computing and Computer-Assisted Intervention

Quantified Score

Hi-index	0.10

Visualization

Abstract

A novel solution to the problem of virtual craniofacial reconstruction using computer vision, graph theory and geometric constraints is proposed. Virtual craniofacial reconstruction is modeled along the lines of the well-known problem of rigid surface registration. The Iterative Closest Point (ICP) algorithm is first employed with the closest set computation performed using the Maximum Cardinality Minimum Weight (MCMW) bipartite graph matching algorithm. Next, the bounding boxes of the fracture surfaces, treated as cycle graphs, are employed to generate multiple candidate solutions based on the concept of graph automorphism. The best candidate solution is selected by exploiting local and global geometric constraints. Finally, the initialization of the ICP algorithm with the best candidate solution is shown to improve surface reconstruction accuracy and speed of convergence. Experimental results on Computed Tomography (CT) scans of real patients are presented.