Removing non-signicant regions in hierarchical clustering and segmentation

Benjamin Perret (1, 2) , Jean Cousty (1, 2) , Silvio Jamil Ferzoli Guimarães (3, 4) , Yukiko Kenmochi (1, 5) , Laurent Najman (1, 2)

1 LIGM - Laboratoire d'Informatique Gaspard-Monge
2 ESIEE Paris
3 PUC Minas - Pontifícia Universidade Católica de Minas Gerais
4 VIPLAB - Audio-visual Information Processing Lab [Sao Gabriel]
5 CNRS - Centre National de la Recherche Scientifique

We propose an efficient algorithm that removes unimportant regions from a hierarchical partition tree, while preserving the hierarchical partition structure. Various experiments demonstrate that applying this algorithm on various classification or segmentation problems does indeed improve the results by a large margin. Code is available online at https://github.com/higra/Higra.

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]; Intelligence artificielle [cs.AI]; Traitement des images [eess.IV]

Benjamin Perret, Jean Cousty, Silvio Jamil Ferzoli Guimarães, Yukiko Kenmochi, Laurent Najman. Removing non-signicant regions in hierarchical clustering and segmentation. Pattern Recognition Letters, 2019, 128, pp.433-439. ⟨10.1016/j.patrec.2019.10.008⟩. ⟨hal-02305469v2⟩