Multiresolution segmentation of natural images: from linear to nonlinear scale-space representations

IEEE Trans Image Process. 2004 Aug;13(8):1104-14. doi: 10.1109/tip.2004.828431.

Abstract

In this paper, we introduce a framework that merges classical ideas borrowed from scale-space and multiresolution segmentation with nonlinear partial differential equations. A non-linear scale-space stack is constructed by means of an appropriate diffusion equation. This stack is analyzed and a tree of coherent segments is constructed based on relationships between different scale layers. Pruning this tree proves to be a very efficient tool for unsupervised segmentation of different classes of images (e.g., natural, medical, etc.). This technique is light on the computational point of view and can be extended to nonscalar data in a straightforward manner.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms*
  • Image Enhancement / methods*
  • Image Interpretation, Computer-Assisted / methods*
  • Linear Models
  • Nonlinear Dynamics
  • Pattern Recognition, Automated*
  • Reproducibility of Results
  • Sensitivity and Specificity
  • Signal Processing, Computer-Assisted*