Featured Works
都市デジタルツイン × AI × XRによる環境設計支援システム
都市デジタルツインと深層学習を統合し、都市環境の認識・可視化・設計支援を高度化する複合現実（XR）システムを開発した。ドローンとARを組み合わせた鳥瞰・一人称視点での3Dオクルージョン処理、GANによる建物除去と未来景観（DR）の生成、さらに実在・非実在都市モデルを用いた建築ファサードのインスタンスセグメンテーション用合成データ自動生成手法を実現している。
本研究は、都市スケールにおける現実と仮想の高精度な統合を可能にし、環境認識および設計意思決定の高度化に寄与する。学術的にも高い被引用実績と上位パーセンタイル評価を獲得しており、メディア掲載や社会実装への展開を通じて、都市計画・防災・インフラ分野への応用が進んでいる。
Key Contributions:
- ドローンARによる都市スケールのオクルージョン処理
- GANによるリアルタイム建物除去と未来景観生成
- 都市デジタルツインを用いた合成データ自動生成
Selected Publications:
- Kikuchi, N., Fukuda, T., Yabuki, N., (2022). Future landscape visualization using a city digital twin: integration of augmented reality and drones with implementation of 3D model-based occlusion handling, Journal of Computational Design and Engineering, 9(2), 837–856, https://doi.org/10.1093/jcde/qwac032
- Zhang, J., Fukuda, T., Yabuki, N., (2022). Automatic generation of synthetic datasets from a city digital twin for use in the instance segmentation of building facades, Journal of Computational Design and Engineering, 9(5), 1737–1755, https://doi.org/10.1093/jcde/qwac086
- Kikuchi, T., Fukuda, T., Yabuki, N., (2023). Development of a synthetic dataset generation method for deep learning of real urban landscapes using a 3D model of a non-existing realistic city, Advanced Engineering Informatics, 58, 102154, https://doi.org/10.1016/j.aei.2023.102154
- 福田知弘. (2024). 1日で学べるXRとメタバース 表現技術検定 公式ガイドブック, フォーラムエイトパブリッシング, https://amzn.asia/d/04NBwAep
深層学習とマルチデータによる都市街路環境の人間中心評価
深層学習、GIS、VR、環境センサを統合し、街路活力・歩行快適性・視覚的快適性の三側面から都市街路環境を人間中心的に評価する手法を提案した。街路空間と人間の行動・認知の関係を多角的に定量化し、実都市（大阪府）において高い有効性と汎用性を実証している。
本研究は、主観的とされてきた「快適性」や「魅力」をデータに基づき定量化することで、都市設計・合意形成への応用を可能にした。関連論文は分野トップジャーナルに掲載され、高い被引用数と上位パーセンタイル評価を獲得している。
Key Contributions:
- 深層学習×GIS×VR×センサの統合評価フレーム
- 街路環境と人間行動・認知の定量化
- 実都市データによる高い汎用性の実証
Selected Publications:
- Li, Y., Yabuki, N., Fukuda, T., (2022). Exploring the association between street built environment and street vitality using deep learning methods, Sustainable Cities and Society, 79, 103656, https://doi.org/10.1016/j.scs.2021.103656
- Li, Y., Yabuki, N., Fukuda, T., (2023). Integrating GIS, deep learning, and environmental sensors for multicriteria evaluation of urban street walkability, Landscape and Urban Planning, 230, 104603, https://doi.org/10.1016/j.landurbplan.2022.104603
- Li, Y., Yabuki, N., Fukuda, T., (2022). Measuring visual walkability perception using panoramic street view images, virtual reality, and deep learning, Sustainable Cities and Society, 86, 104140, https://doi.org/10.1016/j.scs.2022.104140
Urban Digital Twin × AI × XR for Environmental Design Support
This work integrates urban digital twins with deep learning to develop an XR-based system for advanced environmental perception, visualization, and design support. It enables drone-based AR with 3D occlusion handling from both bird’s-eye and first-person perspectives, GAN-based building removal and future landscape generation (DR), and automatic synthetic dataset generation for facade instance segmentation using both real and synthetic urban models.
The system achieves precise integration of real and virtual environments at urban scale, supporting enhanced environmental understanding and design decision-making. The work has received high citation impact and top percentile rankings, and has been widely recognized through media coverage and practical applications in urban planning, disaster mitigation, and infrastructure management.
Key Contributions:
- Drone-based AR with urban-scale occlusion handling
- GAN-based real-time building removal and future visualization
- Synthetic dataset generation using urban digital twins
Selected Publications:
- Kikuchi, N., Fukuda, T., Yabuki, N., (2022). Future landscape visualization using a city digital twin: integration of augmented reality and drones with implementation of 3D model-based occlusion handling, Journal of Computational Design and Engineering, 9(2), 837–856, https://doi.org/10.1093/jcde/qwac032
- Zhang, J., Fukuda, T., Yabuki, N., (2022). Automatic generation of synthetic datasets from a city digital twin for use in the instance segmentation of building facades, Journal of Computational Design and Engineering, 9(5), 1737–1755, https://doi.org/10.1093/jcde/qwac086
- Kikuchi, T., Fukuda, T., Yabuki, N., (2023). Development of a synthetic dataset generation method for deep learning of real urban landscapes using a 3D model of a non-existing realistic city, Advanced Engineering Informatics, 58, 102154, https://doi.org/10.1016/j.aei.2023.102154
- 福田知弘. (2024). 1日で学べるXRとメタバース 表現技術検定 公式ガイドブック, フォーラムエイトパブリッシング, https://amzn.asia/d/04NBwAep
Human-centered Evaluation of Urban Street Environments Using Deep Learning and Multimodal Data
This work proposes a human-centered evaluation framework for urban street environments by integrating deep learning, GIS, VR, and environmental sensors. It quantitatively assesses street vitality, walkability, and visual comfort, capturing the relationship between urban space and human perception and behavior. The approach has been validated in real-world urban settings (Osaka), demonstrating high effectiveness and generalizability.
By quantifying subjective qualities such as comfort and attractiveness, this research enables data-driven urban design and consensus building. The associated publications are ranked in top-tier journals with high citation impact and top percentile metrics.
Key Contributions:
- Integrated framework combining deep learning, GIS, VR, and sensors
- Quantification of human perception and behavior in urban spaces
- Validation using real-world urban data
Selected Publications:
- Li, Y., Yabuki, N., Fukuda, T., (2022). Exploring the association between street built environment and street vitality using deep learning methods, Sustainable Cities and Society, 79, 103656, https://doi.org/10.1016/j.scs.2021.103656
- Li, Y., Yabuki, N., Fukuda, T., (2023). Integrating GIS, deep learning, and environmental sensors for multicriteria evaluation of urban street walkability, Landscape and Urban Planning, 230, 104603, https://doi.org/10.1016/j.landurbplan.2022.104603
- Li, Y., Yabuki, N., Fukuda, T., (2022). Measuring visual walkability perception using panoramic street view images, virtual reality, and deep learning, Sustainable Cities and Society, 86, 104140, https://doi.org/10.1016/j.scs.2022.104140