Rethinking panoptic segmentation in remote sensing : a hybrid approach using semantic segmentation and non-learning methods

Registro completo de metadados
MetadadosDescriçãoIdioma
Autor(es): dc.contributorUniversity of Brasilia, Department of Computer Science-
Autor(es): dc.contributorUniversity of Brasilia, Department of Geography-
Autor(es): dc.contributorUniversity of Brasilia, Department of Geography-
Autor(es): dc.contributorUniversity of Brasilia, Department of Geography-
Autor(es): dc.contributorUniversity of Brasilia, Department of Computer Science-
Autor(es): dc.creatorCarvalho, Osmar Luiz Ferreira de-
Autor(es): dc.creatorCarvalho Júnior, Osmar Abílio de-
Autor(es): dc.creatorAlbuquerque, Anesmar Olino de-
Autor(es): dc.creatorSantana, Níckolas Castro-
Autor(es): dc.creatorBorges, Díbio Leandro-
Data de aceite: dc.date.accessioned2024-10-23T15:46:41Z-
Data de disponibilização: dc.date.available2024-10-23T15:46:41Z-
Data de envio: dc.date.issued2023-09-21-
Data de envio: dc.date.issued2023-09-21-
Data de envio: dc.date.issued2022-05-03-
Fonte completa do material: dc.identifierhttp://repositorio2.unb.br/jspui/handle/10482/46526-
Fonte completa do material: dc.identifierhttps://orcid.org/0000-0002-5619-8525-
Fonte completa do material: dc.identifierhttps://orcid.org/0000-0002-0346-1684-
Fonte completa do material: dc.identifierhttps://orcid.org/0000-0003-1561-7583-
Fonte completa do material: dc.identifierhttps://orcid.org/0000-0001-6133-6753-
Fonte completa do material: dc.identifierhttps://orcid.org/0000-0002-4868-0629-
Fonte: dc.identifier.urihttp://educapes.capes.gov.br/handle/capes/891731-
Descrição: dc.descriptionThis letter proposes a novel method to obtain panoptic predictions by extending the semantic segmentation task with a few non-learning image processing steps, presenting the following benefits: 1) annotations do not require a specific format [e.g., common objects in context (COCO)]; 2) fewer parameters (e.g., single loss function and no need for object detection parameters); and 3) a more straightforward sliding windows implementation for large image classification (still unexplored for panoptic segmentation). Semantic segmentation models do not individualize touching objects, as their predictions can merge; i.e., a single polygon represents many targets. Our method overcomes this problem by isolating the objects using borders on the polygons that may merge. The data preparation requires generating a one-pixel border, and for unique object identification, we create a list with the isolated polygons, attribute a different value to each one, and use the expanding border (EB) algorithm for those with borders. Although any semantic segmentation model applies, we used the U-Net with three backbones (EfficientNet-B5, EfficientNet-B3, and EfficientNet-B0). The results show that the following hold: 1) the EfficientNet-B5 had the best results with 70% mean intersection over union (mIoU); 2) the EB algorithm presented better results for better models; 3) the panoptic metrics show a high capability of identifying things and stuff with 65 panoptic quality (PQ); and 4) the sliding windows on a 2560×2560 -pixel area has shown promising results, in which the ratio of merged objects by correct predictions was lower than 1% for all classes.-
Descrição: dc.descriptionInstituto de Ciências Exatas (IE)-
Descrição: dc.descriptionDepartamento de Ciência da Computação (IE CIC)-
Descrição: dc.descriptionInstituto de Ciências Humanas (ICH)-
Descrição: dc.descriptionDepartamento de Geografia (ICH GEA)-
Idioma: dc.languageen-
Publicador: dc.publisherIEEE-
Relação: dc.relationhttps://ieeexplore.ieee.org/document/9766343-
Direitos: dc.rightsAcesso Restrito-
Palavras-chave: dc.subjectSensoriamento remoto-
Palavras-chave: dc.subjectSegmentação semântica-
Palavras-chave: dc.subjectAprendizagem profunda-
Palavras-chave: dc.subjectSegmentação de imagens-
Palavras-chave: dc.subjectSegmentação panótica-
Título: dc.titleRethinking panoptic segmentation in remote sensing : a hybrid approach using semantic segmentation and non-learning methods-
Tipo de arquivo: dc.typelivro digital-
Aparece nas coleções:Repositório Institucional – UNB

Não existem arquivos associados a este item.