A Sharp Concentration-Based Adaptive Segmentation Algorithm

We propose an adaptive procedure for segmenting images by merging of homogeneous regions. The algorithm is based on sharp concentration inequalities and is tailored to avoid over- and under-merging by controlling simultaneously the type I and II errors in the associated statistical testing problem.

• Type: Journal
• Year: 2010
• Where: ISVC
• Authors: Christophe Fiorio, Andre Mas
• Keywords: Applied Computing, Homogeneous Regions, ISVC 2010, Sharp Concentration Inequalities, Statistical Testing Problem