Few 'Zero Level Set'-Shot Learning of Shape Signed Distance Functions in Feature Space - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2022

Few 'Zero Level Set'-Shot Learning of Shape Signed Distance Functions in Feature Space

Résumé

Unlabelled - Grasslands (natural, semi-natural and improved) occupy approximately one-third of the terrestrial biosphere and are key for global ecosystem service provision, storing up to 30% of soil organic carbon (SOC). To date, most research on soil carbon (C) sequestration has focused on croplands where the levels of native soil organic matter (SOM) are typically low and significant potential exists to replenish SOM stocks. However, with the renewed push to achieve "net zero" C emissions by 2050, grasslands may offer an additional C store, utilising tools such as biochar. Here, we critically evaluate the potential for biochar as a technology for increasing grassland C stocks, identifying a number of practical, economic, social and legislative challenges that need to be addressed before the widescale adoption of biochar may be achieved. We critically assess the current knowledge within the field of grassland biochar research in the context of ecosystem service provision and provide opinions on the applicability of biochar as an amendment to different types of grassland (improved, semi-improved and unimproved) and the potential effect on ecosystem provision using a range of application techniques in the topsoil and subsoil. We concluded that the key question remains, is it possible for managed grasslands to store more C, without causing a loss in additional ecosystem services? To address this question future research must take a more multidisciplinary and holistic approach when evaluating the potential role of biochar at sequestering C in grasslands to mitigate climate change. Supplementary information - The online version contains supplementary material available at 10.1007/s42773-023-00232-y.
Fichier principal
Vignette du fichier
136920556.pdf (4.97 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03941749 , version 1 (16-01-2023)

Identifiants

Citer

Amine Ouasfi, Adnane Boukhayma. Few 'Zero Level Set'-Shot Learning of Shape Signed Distance Functions in Feature Space. ECCV 2022 - European Conference on Computer Vision, Oct 2022, Tel Aviv, Israel. pp.561-578, ⟨10.1007/978-3-031-19824-3_33⟩. ⟨hal-03941749⟩
25 Consultations
30 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More