Evaluation of morphological hierarchies for supervised segmentation

• Main Authors: Perret, Benjamin, Cousty, Jean, Rivera Ura, Jean Carlo, Guimarães, Silvio Jamil F.
• Other Authors: Laboratoire d'Informatique Gaspard-Monge (LIGM), Université Paris-Est Marne-la-Vallée (UPEM)-École des Ponts ParisTech (ENPC)-ESIEE Paris-Fédération de Recherche Bézout (BEZOUT), Centre National de la Recherche Scientifique (CNRS)-Centre National de la Recherche Scientifique (CNRS)-Centre National de la Recherche Scientifique (CNRS), Audio-visual Information Processing Lab Sao Gabriel (VIPLAB), Pontifical Catholic University of Minas Gerais Belo Horizonte, Benediktsson, J.A., Chanussot, J., Najman, L., Talbot, H, ANR-10-BLAN-0205,KIDICO,Intégration des connaissances pour la convolution discrète, la segmentation et la reconstruction d'informations dans les images digitales(2010)
• Format: Conference Object
• Language: English
• Published: HAL CCSD 2015
• Subjects: hierarchy; supervised segmentation; morphology; [INFO.INFO-TI]Computer Science [cs]/Image Processing [eess.IV]; Iceland
• Online Access: https://hal.science/hal-01142072; https://hal.science/hal-01142072/document; https://hal.science/hal-01142072/file/Perret%20-%20ISMM%202015%20-%20Evaluation%20Hierarchies.pdf; https://doi.org/10.1007/978-3-319-18720-4_4

International audience We propose a quantitative evaluation of morphological hierarchies (quasi-flat zones, constraint connectivity, watersheds, observation scale) in a novel framework based on the marked segmentation problem. We created a set of automatically generated markers for the one object image datasets of Grabcut and Weizmann. In order to evaluate the hierarchies , we applied the same segmentation strategy by combining several parameters and markers. Our results, which shows important differences among the considered hierarchies, give clues to understand the behaviour of each method in order to choose the best one for a given application. The code and the marker datasets are available online.