Robust active appearance model matching

Inf Process Med Imaging. 2005:19:114-25. doi: 10.1007/11505730_10.

Abstract

A novel robust active appearance model (AAM) matching algorithm is presented. The method consists of two main stages. First, initial residuals are clustered by a non parametric mean shift mode detection step. Second, modes without gross outliers are selected using an objective function. Robustness of the matching procedure is demonstrated on a variety of examples with different noise conditions. The proposed algorithm outperformed the conventional AAM matching on images with gross disturbances and can tolerate up to 40% of disturbed data.

Publication types

  • Evaluation Study
  • Research Support, N.I.H., Extramural
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms*
  • Artificial Intelligence*
  • Cluster Analysis
  • Computer Simulation
  • Finger Phalanges / diagnostic imaging*
  • Humans
  • Models, Biological
  • Pattern Recognition, Automated / methods*
  • Radiographic Image Enhancement / methods*
  • Radiographic Image Interpretation, Computer-Assisted / methods*
  • Reproducibility of Results
  • Sensitivity and Specificity
  • Subtraction Technique*