% Literature for docs/ml-feasibility.md (research of 2026-11-01).
% DOIs checked via Crossref/OpenAlex/publisher pages. `comment` fields are own summaries, not abstracts.

@InProceedings{Hu2016,
  author    = {Hu, Yitao and Liu, Xiaochen and Nath, Suman and Govindan, Ramesh},
  title     = {{ALPS}: Accurate Landmark Positioning at City Scales},
  booktitle = {Proceedings of the 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp '16)},
  year      = {2016},
  pages     = {1147--1158},
  publisher = {ACM},
  doi       = {10.1145/2971648.2971674},
  comment   = {GSV-based detection+triangulation incl. fire hydrants; LA 90004: 262/330 hydrants (79.4%), median error 4.96 m, 87% within 10 m; 95.5% of visible with Bing Streetside added (figures via X. Liu dissertation, arXiv:2011.00427).}
}

@Article{Krylov2018,
  author  = {Krylov, Vladimir A. and Kenny, Eamonn and Dahyot, Rozenn},
  title   = {Automatic Discovery and Geotagging of Objects from Street View Imagery},
  journal = {Remote Sensing},
  year    = {2018},
  volume  = {10},
  number  = {5},
  pages   = {661},
  doi     = {10.3390/rs10050661},
  comment = {Segmentation + monocular depth + MRF triangulation on GSV; traffic lights P 0.922/R 0.940; poles P 0.960/R 0.935, mean error 0.98 m. Open access.}
}

@InProceedings{Wegner2016,
  author    = {Wegner, Jan D. and Branson, Steve and Hall, David and Schindler, Konrad and Perona, Pietro},
  title     = {Cataloging Public Objects Using Aerial and Street-Level Images -- Urban Trees},
  booktitle = {2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  year      = {2016},
  pages     = {6014--6023},
  doi       = {10.1109/CVPR.2016.647},
  comment   = {Multi-view fusion of aerial and street-level imagery to catalogue public objects (80k trees, Pasadena).}
}

@Article{Boller2019,
  author  = {Boller, Dominik and Moy de Vitry, Matthew and Wegner, Jan D. and Leit{\~a}o, Jo{\~a}o P.},
  title   = {Automated localization of urban drainage infrastructure from public-access street-level images},
  journal = {Urban Water Journal},
  year    = {2019},
  volume  = {16},
  number  = {7},
  pages   = {480--493},
  doi     = {10.1080/1573062X.2019.1687743},
  comment = {Manholes and storm inlets in 5,000 GSV panoramas (Switzerland); conference version: manholes R~80%/P~82%, inlets 76%/80%; localization errors up to several metres.}
}

@Article{Santos2020,
  author  = {dos Santos, Anderson Aparecido and Marcato Junior, Jos{\'e} and de Andrade Silva, Jonathan and Pereira, Rodrigo and Matos, Daniel and Menezes, Geazy and Higa, Leandro and Eltner, Anette and Ramos, Ana Paula and Osco, Lucas and Gon{\c{c}}alves, Wesley},
  title   = {Storm-Drain and Manhole Detection Using the {RetinaNet} Method},
  journal = {Sensors},
  year    = {2020},
  volume  = {20},
  number  = {16},
  pages   = {4450},
  doi     = {10.3390/s20164450},
  comment = {297 car-mounted images; RetinaNet AP50 manhole 94-100%, storm drain 84-89%; dataset released. Open access.}
}

@InProceedings{Timofte2011,
  author    = {Timofte, Radu and Van Gool, Luc},
  title     = {Multi-view manhole detection, recognition, and {3D} localisation},
  booktitle = {2011 IEEE International Conference on Computer Vision Workshops (ICCV Workshops)},
  year      = {2011},
  pages     = {188--195},
  doi       = {10.1109/ICCVW.2011.6130242},
  comment   = {Vision+GPS manhole mapping from vehicle imagery with multi-view 3D localisation.}
}

@Article{Pasquet2016,
  author  = {Pasquet, J{\'e}r{\^o}me and Desert, Thibault and Bartoli, Olivier and Chaumont, Marc and Delenne, Carole and Subsol, G{\'e}rard and Derras, Mustapha and Chahinian, Nan{\'e}e},
  title   = {Detection of Manhole Covers in High-Resolution Aerial Images of Urban Areas by Combining Two Methods},
  journal = {IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing},
  year    = {2016},
  volume  = {9},
  number  = {5},
  pages   = {1802--1807},
  doi     = {10.1109/JSTARS.2015.2504401},
  comment = {4 cm orthophotos; HOG/SVM ~40% of covers at 80% precision; ML requires very high resolution.}
}

@Article{Commandre2017,
  author  = {Commandr{\'e}, B. and En-Nejjary, D. and Pibre, L. and Chaumont, M. and Delenne, C. and Chahinian, N.},
  title   = {Manhole Cover Localization in Aerial Images with a Deep Learning Approach},
  journal = {International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences},
  year    = {2017},
  volume  = {XLII-1/W1},
  pages   = {333--338},
  doi     = {10.5194/isprs-archives-XLII-1-W1-333-2017},
  comment = {CNN sliding window on 5 cm aerial images; >49% of manholes at 75% precision. Open access.}
}

@Article{Liu2019,
  author  = {Liu, Wei and Cheng, Dayu and Yin, Pengcheng and Yang, Mengyuan and Li, Erzhu and Xie, Meng and Zhang, Lianpeng},
  title   = {Small Manhole Cover Detection in Remote Sensing Imagery with Deep Convolutional Neural Networks},
  journal = {ISPRS International Journal of Geo-Information},
  year    = {2019},
  volume  = {8},
  number  = {1},
  pages   = {49},
  doi     = {10.3390/ijgi8010049},
  comment = {5 cm aerial imagery, 23,252 covers; best P 0.864/R 0.927/F1 0.895 vs YOLOv3 F1 0.74. Open access.}
}

@Article{Yu2015,
  author  = {Yu, Yongtao and Guan, Haiyan and Ji, Zheng},
  title   = {Automated Detection of Urban Road Manhole Covers Using Mobile Laser Scanning Data},
  journal = {IEEE Transactions on Intelligent Transportation Systems},
  year    = {2015},
  volume  = {16},
  number  = {6},
  pages   = {3258--3269},
  doi     = {10.1109/TITS.2015.2413812},
  comment = {Mobile laser scanning; completeness 0.955, correctness 0.959, F1 0.957.}
}

@Article{Mattheuwsen2020,
  author  = {Mattheuwsen, Lukas and Vergauwen, Maarten},
  title   = {Manhole Cover Detection on Rasterized Mobile Mapping Point Cloud Data Using Transfer Learned Fully Convolutional Neural Networks},
  journal = {Remote Sensing},
  year    = {2020},
  volume  = {12},
  number  = {22},
  pages   = {3820},
  doi     = {10.3390/rs12223820},
  comment = {Rasterised MMS point clouds, VGG-16 FCN; recall/precision/F2 0.973; 95% horizontal CI 16.5 cm. Open access.}
}

@Article{Ai2016,
  author  = {Ai, Chengbo and Tsai, Yichang James},
  title   = {Automated Sidewalk Assessment Method for Americans with Disabilities Act Compliance Using Three-Dimensional Mobile Lidar},
  journal = {Transportation Research Record: Journal of the Transportation Research Board},
  year    = {2016},
  volume  = {2542},
  number  = {1},
  pages   = {25--32},
  doi     = {10.3141/2542-04},
  comment = {Mobile lidar + video: sidewalk width, cross slope, curb ramps; validated against field survey.}
}

@Article{Hosseini2023,
  author  = {Hosseini, Maryam and Sevtsuk, Andres and Miranda, Fabio and Cesar Jr, Roberto M. and Silva, Claudio T.},
  title   = {Mapping the walk: A scalable computer vision approach for generating sidewalk network datasets from aerial imagery},
  journal = {Computers, Environment and Urban Systems},
  year    = {2023},
  volume  = {101},
  pages   = {101950},
  doi     = {10.1016/j.compenvurbsys.2023.101950},
  comment = {Tile2Net, ~0.19 m/px; sidewalk IoU 82.67% (P 0.90, R 0.92). Code: github.com/VIDA-NYU/tile2net.}
}

@Article{Ning2022,
  author  = {Ning, Huan and Ye, Xinyue and Chen, Zhihui and Liu, Tao and Cao, Tianzhi},
  title   = {Sidewalk extraction using aerial and street view images},
  journal = {Environment and Planning B: Urban Analytics and City Science},
  year    = {2022},
  volume  = {49},
  number  = {1},
  pages   = {7--22},
  doi     = {10.1177/2399808321995817},
  comment = {Aerial segmentation + street-view gap filling; P/R ~0.9 in four US counties.}
}

@Article{Lieu2026,
  author  = {Lieu, Seung Jae and Guhathakurta, Subhrajit},
  title   = {A novel approach for estimating sidewalk width from street view images and computer vision},
  journal = {Environment and Planning B: Urban Analytics and City Science},
  year    = {2026},
  volume  = {53},
  number  = {5},
  pages   = {1129--1144},
  doi     = {10.1177/23998083251369602},
  comment = {Street view at two pitch angles, segmentation + trigonometry for sidewalk width.}
}

@Misc{Tan2026,
  author        = {Tan, Kaizhen and Zhang, Fan},
  title         = {{UrbanVGGT}: Scalable Sidewalk Width Estimation from Street View Images},
  year          = {2026},
  eprint        = {2603.22531},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  doi           = {10.48550/arXiv.2603.22531},
  note          = {Preprint, not peer-reviewed},
  comment       = {Single-image metric sidewalk width; MAE 0.252 m, 95.5% within 0.5 m (Washington DC).}
}

@InProceedings{Neuhold2017,
  author    = {Neuhold, Gerhard and Ollmann, Tobias and Rota Bul{\`o}, Samuel and Kontschieder, Peter},
  title     = {The {Mapillary Vistas} Dataset for Semantic Understanding of Street Scenes},
  booktitle = {2017 IEEE International Conference on Computer Vision (ICCV)},
  year      = {2017},
  pages     = {5000--5009},
  doi       = {10.1109/ICCV.2017.534},
  comment   = {v1.2: Fire Hydrant, Manhole, Catch Basin, Curb, Sidewalk; v2.0 (124 classes) adds Water Valve.}
}

@InProceedings{Ertler2020,
  author    = {Ertler, Christian and Mislej, Jerneja and Ollmann, Tobias and Porzi, Lorenzo and Neuhold, Gerhard and Kuang, Yubin},
  title     = {The {Mapillary} Traffic Sign Dataset for Detection and Classification on a Global Scale},
  booktitle = {Computer Vision -- ECCV 2020},
  series    = {Lecture Notes in Computer Science},
  year      = {2020},
  pages     = {68--84},
  publisher = {Springer},
  doi       = {10.1007/978-3-030-58592-1_5},
  comment   = {Global traffic-sign dataset; pretraining for a DIN 4066 sign detector.}
}

@InProceedings{Lin2014,
  author    = {Lin, Tsung-Yi and Maire, Michael and Belongie, Serge and Hays, James and Perona, Pietro and Ramanan, Deva and Doll{\'a}r, Piotr and Zitnick, C. Lawrence},
  title     = {Microsoft {COCO}: Common Objects in Context},
  booktitle = {Computer Vision -- ECCV 2014},
  series    = {Lecture Notes in Computer Science},
  year      = {2014},
  pages     = {740--755},
  publisher = {Springer},
  doi       = {10.1007/978-3-319-10602-1_48},
  comment   = {Includes the 'fire hydrant' class (above-ground, mostly US-style).}
}

@Article{Yang2022,
  author  = {Yang, Zhehui and Zhao, Chenbo and Maeda, Hiroya and Sekimoto, Yoshihide},
  title   = {Development of a Large-Scale Roadside Facility Detection Model Based on the {Mapillary} Dataset},
  journal = {Sensors},
  year    = {2022},
  volume  = {22},
  number  = {24},
  pages   = {9992},
  doi     = {10.3390/s22249992},
  comment = {Vistas v2, 13 roadside facility classes; YOLOv7 P 87.57%, R 72.60%, F1 78.70%. Open access.}
}

@InProceedings{Rahul2019,
  author    = {Rahul, Rohit and Paliwal, Shubham Singh and Sharma, Monika and Vig, Lovekesh},
  title     = {Automatic Information Extraction from Piping and Instrumentation Diagrams},
  booktitle = {Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods (ICPRAM)},
  year      = {2019},
  pages     = {163--172},
  publisher = {SciTePress},
  doi       = {10.5220/0007376401630172},
  comment   = {Symbol segmentation + OCR + line tracing for scanned P&IDs; closest analogue for scanned network plans.}
}

@InProceedings{Ignjatic2018,
  author    = {Ignjati{\'c}, Jelena and Nikoli{\'c}, Bojana and Rikalovi{\'c}, Aleksandar},
  title     = {Deep learning for historical cadastral maps digitization: overview, challenges and potential},
  booktitle = {WSCG 2018: Poster Papers Proceedings},
  series    = {Computer Science Research Notes},
  number    = {2803},
  year      = {2018},
  pages     = {42--47},
  doi       = {10.24132/CSRN.2018.2803.6},
  comment   = {Overview of deep learning for vectorising scanned cadastral maps. Open access.}
}