Please use this identifier to cite or link to this item:
|Title:||Constrained Connectivity for Hierarchical Image Segmentation and Simplification|
|Citation:||IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE vol. 30 no. 7 p. 1132-1145|
|Publisher:||IEEE COMPUTER SOC|
|Type:||Articles in Journals|
|Abstract:||This paper introduces a series of new connectivity relations whose equivalent classes (called connected components) define unique partitions of the definition domain of a given grey tone image. The primary relation relies on two range parameters controlling the intensity variations occurring within each connected component. The first range parameter restricts the intensity differences between successive pixels of the connected paths linking arbitrary pairs of pixels of the connected component. The second range parameter limits the difference between the maximum and minimum intensity values of each connected component. By introducing a connectivity index measuring the degree of tightness of each connected component, further useful connectivity relations are suggested. Fine to coarse partition hierarchies are obtained by increasing the values of the range and connectivity index parameters. The paper also includes a generalisation to multichannel images, pseudo-code for an implementation based on queue and stack data structures, applications to image processing tasks such as segmentation and image simplification, and a comparison with a variety of related techniques.|
|JRC Institute:||Institute for Environment and Sustainability|
Files in This Item:
There are no files associated with this item.
Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.