Pré-Publication, Document De Travail Année : 2025

Evaluation of Polarimetric Fusion for Semantic Segmentation in Aquatic Environments

Tom Bourbon
  • Fonction : Auteur
  • PersonId : 1588128
Cedric Pradalier

Résumé

Accurate segmentation of floating debris on water is often compromised by surface glare and changing outdoor illumination. Polarimetric imaging offers a single-sensor route to mitigate water-surface glare that disrupts semantic segmentation of floating objects. We benchmark state-of-the-art fusion networks on PoTATO, a public dataset of polarimetric images of plastic bottles in inland waterways, and compare their performance with single-image baselines using traditional models. Our results indicate that polarimetric cues help recover low-contrast objects and suppress reflection-induced false positives, raising mean IoU and lowering contour error relative to RGB inputs. These sharper masks come at a cost: the additional channels enlarge the models increasing the computational load and introducing the risk of new false positives. By providing a reproducible, diagnostic benchmark and publicly available code, we hope to help researchers choose if polarized cameras are suitable for their applications and to accelerate related research.

Fichier principal
Vignette du fichier
Polarimetric_Segmentation___VCIP25.pdf (7.16 Mo) Télécharger le fichier

Dates et versions

hal-05288763 , version 1 (29-09-2025)

Licence

Identifiants

  • HAL Id : hal-05288763 , version 1

Citer

Luis Felipe Wolf Batista, Tom Bourbon, Cedric Pradalier. Evaluation of Polarimetric Fusion for Semantic Segmentation in Aquatic Environments. 2025. ⟨hal-05288763⟩
81 Consultations
148 Téléchargements

Partager

  • More