On power Jaccard losses for semantic segmentation - Archive ouverte HAL Access content directly
Conference Papers Year :

On power Jaccard losses for semantic segmentation

Jean-Emmanuel Deschaud
Francois Goulette
Andrés Serna

Abstract

In this work, a new generalized loss function is proposed called power Jaccard to perform semantic segmentation tasks. It is compared with classical loss functions in different scenarios, including gray level and color image segmentation, as well as 3D point cloud segmentation. The results show improved performance, stability and convergence. We made available the code with our proposal with a demonstrative example.
Fichier principal
Vignette du fichier
On_Power_losses_for_semantic_segmentation.pdf (1.02 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03139997 , version 1 (12-02-2021)

Identifiers

  • HAL Id : hal-03139997 , version 1

Cite

David Duque-Arias, Santiago Velasco-Forero, Jean-Emmanuel Deschaud, Francois Goulette, Andrés Serna, et al.. On power Jaccard losses for semantic segmentation. VISAPP 2021 : 16th International Conference on Computer Vision Theory and Applications, Feb 2021, Vienne (on line), Austria. ⟨hal-03139997⟩
4941 View
854 Download

Share

Gmail Facebook Twitter LinkedIn More