A framework for multi-segmentation natural images

• Grant number: 12/17439-2
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: November 01, 2012
• End date: October 31, 2013
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Paulo Sergio Silva Rodrigues
• Grantee: Lukercio de Abreu Lopes
• Host Institution: Centro Universitário FEI (UNIFEI). Campus de São Bernardo do Campo. São Bernardo do Campo , SP, Brazil

Abstract

Methods of image segmentation based on entropy have been gaining increasing attention from researchers, mainly involving the search for automatic thresholds. A digital image can be thresholded with one or more threshold values, but the amount of thresholds is still a point of controversy among researchers. Another issue that is widely discussed in the area concerns the quality of the segmentation, which is crucially dependent on human subjectivity. This project aims to advance these two problems and presents methods based on meta-heuristics that underline entropy for a more efficient calculation of the optimal threshold. Also, alternatively presents a human segmented database, which serves as a reference ground truth for the quality of the responses. We use parallel computation for exhaustive methods that serve to the upper bound efficient response.(AU)