D4SC: Deep Supervised Semantic Segmentation for Seabed Characterisation in Low-Label Regime

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/CongresprocedureConferentiebijdragepeer review

3 Downloads (Pure)

Samenvatting

Seabed characterisation consists in the study of the physical and biological properties of the bottom of the oceans. It is effectively achieved with sonar, a remote sensing method that captures acoustic backscatter of the seabed. Classical Machine Learning (ML) and Deep Learning (DL) research have failed to successfully address the automatic mapping of the seabed from noisy sonar data. This work introduces the Deep Supervised Semantic Segmentation model for Seabed Characterisation (D4SC), a novel U-Net-like model tailored to such data and low-label regime, and proposes a new end-to-end processing pipeline for seabed semantic segmentation. That dual contribution achieves state-of-the-art results on a high resolution Synthetic Aperture Sonar (SAS) survey dataset.

Originele taal-2Engels
TitelIGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium, Proceedings
UitgeverijInstitute of Electrical and Electronics Engineers Inc.
Pagina's6932-6935
Aantal pagina's4
ISBN van elektronische versie9798350320107
DOI's
StatusGepubliceerd - 2023
Evenement2023 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2023 - Pasadena, Verenigde Staten van Amerika
Duur: 16 jul. 202321 jul. 2023
https://2023.ieeeigarss.org/

Publicatie series

NaamInternational Geoscience and Remote Sensing Symposium (IGARSS)
Volume2023-July

Congres

Congres2023 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2023
Verkorte titelIGARSS 2023
Land/RegioVerenigde Staten van Amerika
StadPasadena
Periode16/07/2321/07/23
Internet adres

Vingerafdruk

Duik in de onderzoeksthema's van 'D4SC: Deep Supervised Semantic Segmentation for Seabed Characterisation in Low-Label Regime'. Samen vormen ze een unieke vingerafdruk.

Citeer dit