Centre de documentation
scientifique

Measuring visual walkability perception using panoramic street view images, virtual reality, and deep learning / Yunqin Li in Sustainable Cities and Society, vol 86 (November 2022)  

Public [article]  inSustainable Cities and Society > vol 86 (November 2022) . - n° 104140 Titre : Measuring visual walkability perception using panoramic street view images, virtual reality, and deep learning Type de document : Article/Communication Auteurs : Yunqin Li, Auteur ; Nobuyoshi Yabuki, Auteur ; Tomohiro Fukuda, Auteur Année de publication : 2022 Article en page(s) : n° 104140 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] apprentissage profond [Termes IGN] classification par séparateurs à vaste marge [Termes IGN] corrélation [Termes IGN] image panoramique [Termes IGN] image Streetview [Termes IGN] modèle de régression [Termes IGN] piéton [Termes IGN] réalité virtuelle [Termes IGN] scène urbaine [Termes IGN] segmentation sémantique [Termes IGN] vision Résumé : (auteur) Measuring perceptions of visual walkability in urban streets and exploring the associations between the visual features of the street built environment that make walking attractive to humans are both theoretically and practically important. Previous studies have used either environmental audits and subjective evaluations that have limitations in terms of cost, time, and measurement scale, or computer-aided audits based on natural street view images (SVIs) but with gaps in real perception. In this study, a virtual reality panoramic image-based deep learning framework is proposed for measuring visual walkability perception (VWP) and then quantifying and visualizing the contributing visual features. A VWP classification deep multitask learning (VWPCL) model was first developed and trained on human ratings of panoramic SVIs in virtual reality to predict VWP in six categories. Second, a regression model was used to determine the degree of correlation of various objects with one of the six VWP categories based on semantic segmentation. Furthermore, an interpretable deep learning model was used to assist in identifying and visualizing elements that contribute to VWP. The experiment validated the accuracy of the VWPCL model for predicting VWP. The results represent a further step in understanding the interplay of VWP and street-level semantics and features. Numéro de notice : A2022-816 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.1016/j.scs.2022.104140 Date de publication en ligne : 21/08/2022 En ligne : https://doi.org/10.1016/j.scs.2022.104140 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=101982 [article]

Exploring the association between street built environment and street vitality using deep learning methods / Yunqin Li in Sustainable Cities and Society, vol 79 (April 2022)  

Public [article]  inSustainable Cities and Society > vol 79 (April 2022) . - n° 103656 Titre : Exploring the association between street built environment and street vitality using deep learning methods Type de document : Article/Communication Auteurs : Yunqin Li, Auteur ; Nobuyoshi Yabuki, Auteur ; Tomohiro Fukuda, Auteur Année de publication : 2022 Article en page(s) : n° 103656 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications SIG [Termes IGN] apprentissage profond [Termes IGN] attractivité (aménagement) [Termes IGN] bati [Termes IGN] image Streetview [Termes IGN] Japon [Termes IGN] morphologie urbaine [Termes IGN] OpenStreetMap [Termes IGN] piéton [Termes IGN] planification urbaine [Termes IGN] processus de hiérarchisation analytique [Termes IGN] régression linéaire [Termes IGN] scène urbaine [Termes IGN] segmentation sémantique [Termes IGN] système d'information géographique [Termes IGN] urbanisme [Termes IGN] ville intelligente Résumé : (auteur) Street vitality has become an essential indicator for evaluating the attractiveness and potential of the sustainable development of urban blocks, and it can be reflected by the type and the frequency of people's pedestrian activities on the street. While it is recognized that street built environment features affect pedestrian behavior and street vitality, quantifying the impact of these characteristics remains inconclusive. This paper proposes an automated deep learning approach to quantitatively explore the association between the street built environment and street vitality. First, we established a deep learning model for street vitality classification for automatic evaluation of street vitality based on the volumes and activities of pedestrians in the street through multiple object tracking and scene classification. Then, we applied semantic segmentation to measure five selected vitality-related street built environment variables. Finally, a linear regression model was applied to evaluate the built environment variables’ significance and effects on street vitality. To verify our method's accuracy and applicability, we selected a commercial complex in Osaka as an illustrative example. The experimental results highlight that street width and transparency have significant positive effects on street vitality. Compared with traditional methods, our approach is feasible, reliable, transferable, and more efficient. Numéro de notice : A2022-266 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE/URBANISME Nature : Article DOI : 10.1016/j.scs.2021.103656 Date de publication en ligne : 10/01/2022 En ligne : https://doi.org/10.1016/j.scs.2021.103656 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100271 [article]