Detection of Arterial Calcification in Mammograms by Random Walks

Authors:
Jie-Zhi Cheng;Elodia B. Cole;Etta D. Pisano;Dinggang Shen
Affiliations:
Biomedical Research Imaging Center, University of North Carolina at Chapel Hill, Chapel Hill, USA NC 27599;Biomedical Research Imaging Center, University of North Carolina at Chapel Hill, Chapel Hill, USA NC 27599;Biomedical Research Imaging Center, University of North Carolina at Chapel Hill, Chapel Hill, USA NC 27599;Biomedical Research Imaging Center, University of North Carolina at Chapel Hill, Chapel Hill, USA NC 27599
Venue:
IPMI '09 Proceedings of the 21st International Conference on Information Processing in Medical Imaging
Year:
2009

Citing 5
Cited 1

On sequential Monte Carlo sampling methods for Bayesian filtering

Statistics and Computing
Muliscale Vessel Enhancement Filtering

MICCAI '98 Proceedings of the First International Conference on Medical Image Computing and Computer-Assisted Intervention
Learning Probabilistic Models for Contour Completion in Natural Images

International Journal of Computer Vision
Bayesian tracking of elongated structures in 3D images

IPMI'07 Proceedings of the 20th international conference on Information processing in medical imaging
Joint sulci detection using graphical models and boosted priors

IPMI'07 Proceedings of the 20th international conference on Information processing in medical imaging

Quantification of vascular calcifications on digitized mammograms

IWDM'10 Proceedings of the 10th international conference on Digital Mammography

Quantified Score

Hi-index	0.00

Visualization

Abstract

A fully automatic algorithm is developed for breast arterial calcification extraction in mammograms. This algorithm is implemented in two major steps: a random-walk based tracking step and a compiling and linking step. With given seeds from detected calcification points, the tracking algorithm traverses the vesselness map by exploring the uncertainties of three tracking factors, i.e., traversing direction, jumping distance, and vesselness value, to generate all possible sampling paths. The compiling and linking algorithm further organizes and groups all sampling paths into calcified vessel tracts. The experimental results show that the performance of the proposed automatic calcification extraction algorithm is statistically close to that obtained by manual delineations.