東京大学大学院 情報理工学系研究科の授業『知能情報論』2024年第10回『三次元再構成』の講義資料です。
1. 三次元再構成とは
2. 3D表現,特にニューラル場
3. ニューラル場の最適化
4. 三次元再構成の学習
5. 発展的な話題
6. まとめ
リンク一覧
- p5. [Mildenhall+ 2020] https://www.matthewtancik.com/nerf
- p6. 首里城デジタル復元 https://www.our-shurijo.org/
- p6. [Tesla AI Day 2023] https://www.youtube.com/watch?v=ODSJsviD_SU
- p10. [Watercolor by Shibasaki] https://www.youtube.com/watch?v=WWq4PoM9jXw
- p16. [日本視覚学会, 2022] https://www.agu.ac.jp/~sakano/disparity.html
- p18. [Ranftl+ 2021] https://arxiv.org/abs/2103.13413
- p18. [Mescheder+ 2019] https://github.com/autonomousvision/occupancy_networks
- p19. [CROSS SPACE] https://vision.xspace.tokyo/3dcat/
- p19. [Julian Beever] https://www.julianbeever.net/
- p23. URL http://www.bilderzucht.de/blog/3d-pixel-voxel/
- p24. URL https://doc.cgal.org/latest/Orthtree/index.html
- p25. URL https://www.researchgate.net/figure/Point-representation-of-the-Stanford-bunny-model_fig2_224378716
- p26. URL http://www.gradientspace.com/tutorials/2017/8/30/mesh-simplification
- p27. URL https://arxiv.org/abs/1906.06751
- p28. [Mescheder+ 2019] https://arxiv.org/abs/1812.03828
- p30. [Oechsle+ 2019] https://arxiv.org/abs/1905.07259
- p30. [Srinivasan+ 2021] https://pratulsrinivasan.github.io/nerv/
- p30. [Niemeyer+ 2019] https://avg.is.mpg.de/publications/niemeyer2019iccv
- p32. [Mildenhall+ 2020] https://www.matthewtancik.com/nerf
- p32. [Xie+ 2021] https://arxiv.org/abs/2111.11426
- p33. [Chen & Wang 2024] https://arxiv.org/abs/2401.03890
- p34. [URL] https://lumalabs.ai/embed/6af6e301-1267-4c53-89e1-241a4cfd1b80?mode=sparkles&background=%23ffffff&color=%23000000&showTitle=true&loadBg=true&logoPosition=bottom-left&infoPosition=bottom-right&cinematicVideo=undefined&showMenu=false
- p35. Luma AI Featured Captures https://lumalabs.ai/featured
- p37. [Mescheder+ 2019] https://arxiv.org/abs/1812.03828
- p40. [Mildenhall+ 2020] https://www.matthewtancik.com/nerf
- p42. [Mildenhall+ 2020] https://www.matthewtancik.com/nerf
- p43. [Mildenhall+ 2020] https://www.matthewtancik.com/nerf
- p46. SSII 2020 https://docs.google.com/presentation/d/1nE4kiVJK3SMag93VBrUnt8vk5Pg18qYRMa4VS1yhYI4/edit#slide=id.p
- p46. VC 2021 https://docs.google.com/presentation/d/1YypYQHf4BhEAj8BmEgqCNx2NtQrVpOPF4osqvlL6_Bs/edit?usp=sharing
- p46. CVIM 2022-03 https://docs.google.com/presentation/d/1nbyUBucCTFP2-sVLFeo2gP76L3EJTImMFXLmV7OifxI/edit
- p46. CVIMチュートリアル(1) https://www.amazon.co.jp/dp/4320126017
- p47. [Mildenhall+ 2020] https://www.matthewtancik.com/nerf
- p48. [Mildenhall+ 2020] https://www.matthewtancik.com/nerf
- p49. [Mildenhall+ 2020] https://www.matthewtancik.com/nerf
- p50. [Li+ 2021] https://www.cs.cornell.edu/~zl548/NSFF/
- p51. [Martin-Brualla+ 2021] https://nerf-w.github.io/
- p52. [Tancik+ 2022] https://waymo.com/research/block-nerf/
- p53. [Yen-Chen+ 2021] https://yenchenlin.me/inerf/
- p54. [Gao+ 2020] https://arxiv.org/abs/2011.01437
- p55. [Wang+ 2021] https://lingjie0206.github.io/papers/NeuS/index.htm
- p55. [Oechsle+ 2021] https://arxiv.org/abs/2104.10078
- p55. [Yariv+ 2021] https://arxiv.org/abs/2106.12052
- p56. [Mildenhall+ 2020] https://www.matthewtancik.com/nerf
- p57. [Reiser+ 2021] https://github.com/creiser/kilonerf/
- p58. [Müller+ 2022] https://nvlabs.github.io/instant-ngp/
- p59. [Chabra+ 2020] https://arxiv.org/abs/2003.10983
- p60. [Yu+ 2022] https://alexyu.net/plenoxels/
- p61. [URL] https://towardsdatascience.com/a-comprehensive-overview-of-gaussian-splatting-e7d570081362
- p65. [Mescheder+ 2019] https://github.com/autonomousvision/occupancy_networks
- p66. [Kato & Harada 2019] https://arxiv.org/abs/1811.10719
- p68. URL http://www.cs.mun.ca/~omeruvia/philosophy/WireframeBunny.html
- p70. [Xiang+ 2014] https://cvgl.stanford.edu/projects/pascal3d.html
- p73. [Loper+ 2015] https://smpl.is.tue.mpg.de/
- p73. [Romero+ 2017] https://mano.is.tue.mpg.de/
- p74. [Pavlakos+ 2018] https://arxiv.org/abs/1808.02651
- p74. [Lassner+ 2017] https://arxiv.org/abs/1701.02468
- p74. [Wang+ 2020] https://arxiv.org/abs/2003.10873
- p75. [Baek+ 2019] https://arxiv.org/abs/1904.04196
- p75. [Zhang+ 2019] https://arxiv.org/abs/1902.09305
- p75. [Zimmermann+ 2019] https://arxiv.org/abs/1909.04349
- p76. [Nash+ 2020] https://arxiv.org/abs/2002.10880
- p76. [Groueix+ 2018] https://arxiv.org/abs/1802.05384
- p76. [Kato+ 2018] https://github.com/hiroharu-kato/neural_renderer
- p76. [Wang+ 2018] https://arxiv.org/abs/1804.01654
- p76. [Chen+ 2019] https://nv-tlabs.github.io/DIB-R/
- p76. [Kato+ 2018] https://github.com/hiroharu-kato/neural_renderer
- p76. [Kanazawa+ 2018] https://akanazawa.github.io/cmr/
- p77. [Niemeyer+ 2020] https://arxiv.org/abs/1912.07372
- p78. [Kato & Harada 2019] https://arxiv.org/abs/1811.10719
- p80. [Liu+ 2023] https://zero123.cs.columbia.edu/
- p84. 『Stable Diffusionで同じ顔の人物の画像を生成する方法』 https://bocek.co.jp/media/exercise/stable-diffusion/6075/
- p84. 『AIと3Dを利用したアニメ制作 統一性のある背景を様々なアングルから生成』 https://note.com/abubu_nounanka/n/nb5d60e9fc63f
- p84. 『構図の調整に使えるカメラアングルなどのワードまとめ』 https://blogcake.net/ai-angle/
- p87. [Or-El+ 2022] https://stylesdf.github.io/
- p87. [Wang+ 2023] https://ml.cs.tsinghua.edu.cn/prolificdreamer/
- p89-90. [Nguyen-Phuoc+ 2019] https://www.monkeyoverflow.com/hologan-unsupervised-learning-of-3d-representations-from-natural-images/
- p91. [Schwarz+ 2020] https://github.com/autonomousvision/graf
- p92. [Chan+ 2022] https://nvlabs.github.io/eg3d/
- p93. [Sargent+ 2023] https://kylesargent.github.io/vq3d
- p93. [Or-El+ 2022] https://stylesdf.github.io/
- p93. [Deng+ 2009] https://www.image-net.org/
- p94. [Kato & Harada 2019] https://arxiv.org/abs/1811.10719
- p95. [Sargent+ 2023] https://kylesargent.github.io/vq3d
- p96. [Poole 2023] https://dreamfusion3d.github.io/
- p97. [Poole 2023] https://dreamfusion3d.github.io/
- p97. [Wang+ 2023年5月] https://ml.cs.tsinghua.edu.cn/prolificdreamer/
- p98. [Shi+ 2024] https://mv-dream.github.io/
- p99. [Kobayashi+ 2022] https://pfnet-research.github.io/distilled-feature-fields/
- p100. [Kerr+ 2023] https://www.lerf.io/
- p101. SSIIチュートリアル『生成AIと3次元ビジョン』 https://speakerdeck.com/ssii/ssii2024-ts1-ikehata
- p105. https://hiroharu-kato.com/ https://hiroharu-kato.com/
- p105. 三次元構造を考慮した画像生成 https://docs.google.com/presentation/d/1eQUf-fKFj1o3lGxPGFxyqHZTYAl9k85XRg2BPE2F9nc/edit#slide=id.p
- p105. ニューラル3D表現の最新動向 https://docs.google.com/presentation/d/135VTaXJaESu4rDRLe_WbvSe4zq1__dFxBYI_7H3EjBI/edit#slide=id.g12d4f17af70_0_0
- p105. 微分可能レンダリング https://docs.google.com/presentation/d/1nbyUBucCTFP2-sVLFeo2gP76L3EJTImMFXLmV7OifxI/edit#slide=id.p
- p105. モダリティ変換と画像生成 https://www.slideshare.net/100001653434308/ssii-os2