Image segmentation by the image foresting transform

Author(s)

Paulo Andre Vechiatto de Miranda

• Document type: Master's Dissertation
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Instituto de Computação
• Defense date: 2006-02-20

Examining board members

Alexandre Xavier Falcão; Antonio Alberto Fernandes de Oliveira; Siome Klein Goldenstein

• Advisor: Alexandre Xavier Falcão

Abstract

Despite of the progress over the last two decades, image segmentation remains one of the most difficult challenges in image analysis. More specifically, accurate segmentation methods are likely to be complex, computationally expensive, application-dependent, and often require considerable human interaction. We have investigated methods for interactive segmentation that reduce user intervention to simple selection of a few pixels in the image. We are interested in minimizing the user involvement such that automation becomes feasible for some applications where these pixels may be automatically identified. The adopted methodology uses the framework of the image foresting transform (IFT) - a general tool for the design, implementation, and evaluation of image processing operators based on connectivity. The IFT reduces image processing problems into a minimum-cost path forest problem in a graph derived from the image. In this work, we present three new segmentation methods based on the IFT (AU)