請用此 Handle URI 來引用此文件：
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/99656完整後設資料紀錄
| DC 欄位 | 值 | 語言 |
|---|---|---|
| dc.contributor.advisor | 洪一平 | zh_TW |
| dc.contributor.advisor | Yi-Ping Hung | en |
| dc.contributor.author | 吳浩平 | zh_TW |
| dc.contributor.author | Hao-Ping Wu | en |
| dc.date.accessioned | 2025-09-17T16:16:56Z | - |
| dc.date.available | 2025-09-18 | - |
| dc.date.copyright | 2025-09-17 | - |
| dc.date.issued | 2025 | - |
| dc.date.submitted | 2025-08-06 | - |
| dc.identifier.citation | [1] B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoorthi, and R. Ng, “NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis,” in ECCV, 2020.
[2] B. Kerbl, G. Kopanas, T. Leimkühler, and G. Drettakis, “3D Gaussian Splatting for Real-Time Radiance Field Rendering,” ACM Transactions on Graphics, vol. 42, no. 4, July 2023. [Online]. Available: https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/ [3] G. Wu, T. Yi, J. Fang, L. Xie, X. Zhang, W. Wei, W. Liu, Q. Tian, and X. Wang, “4D Gaussian Splatting for Real-Time Dynamic Scene Rendering,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2024, pp. 20310–20320. [4] P. Wang, Z. Zhang, L. Wang, K. Yao, S. Xie, J. Yu, M. Wu, and L. Xu, “V^3: Viewing Volumetric Videos on Mobiles via Streamable 2D Dynamic Gaussians,” ACM Transactions on Graphics (TOG), vol. 43, no. 6, pp. 1–13, 2024. [5] S. Lin, A. Ryabtsev, S. Sengupta, B. Curless, S. Seitz, and I. Kemelmacher Shlizerman, “Real-Time High-Resolution Background Matting,” 2020, arXiv preprint arXiv:2012.07810. [6] Z. Yang, H. Yang, Z. Pan, and L. Zhang, “Real-time Photorealistic Dynamic Scene Representation and Rendering with 4D Gaussian Splatting,” in International Conference on Learning Representations (ICLR), 2024. [7] Z. Xu, Y. Xu, Z. Yu, S. Peng, J. Sun, H. Bao, and X. Zhou, “Representing Long Volumetric Video with Temporal Gaussian Hierarchy,” ACM Transactions on Graphics, vol. 43, no. 6, November 2024. [Online]. Available: https://zju3dv.github.io/longvolcap [8] A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y. Lo, P. Dollár, and R. Girshick, “Segment Anything,” arXiv:2304.02643, 2023. [9] N. Ravi, V. Gabeur, Y.-T. Hu, R. Hu, C. Ryali, T. Ma, H. Khedr, R. Rädle, C. Rolland, L. Gustafson, E. Mintun, J. Pan, K. V. Alwala, N. Carion, C.-Y. Wu, R. Girshick, P. Dollár, and C. Feichtenhofer, “SAM 2: Segment Anything in Images and Videos,” arXiv preprint arXiv:2408.00714, 2024. [Online]. Available: https://arxiv.org/abs/2408.00714 [10] Z. Ke, J. Sun, K. Li, Q. Yan, and R. W. Lau, “MODNet: Real-Time Trimap-Free Portrait Matting via Objective Decomposition,” in AAAI, 2022. [11] S. Lin, L. Yang, I. Saleemi, and S. Sengupta, “Robust High-Resolution Video Matting with Temporal Guidance,” 2021, arXiv preprint arXiv:2108.11515. [12] J. L. Schönberger and J.-M. Frahm, “Structure-from-Motion Revisited,” in Conference on Computer Vision and Pattern Recognition (CVPR), 2016. [13] J. L. Schönberger, E. Zheng, M. Pollefeys, and J.-M. Frahm, “Pixelwise View Selection for Unstructured Multi-View Stereo,” in European Conference on Computer Vision (ECCV), 2016. [14] Y. Wang, Q. Han, M. Habermann, K. Daniilidis, C. Theobalt, and L. Liu, “NeuS2: Fast Learning of Neural Implicit Surfaces for Multi-view Reconstruction,” in Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023. [15] RunPod, “RunPod: Serverless GPU Cloud,” 2025. [Online]. Available: https://www.runpod.io/ [16] A. Pranckevičius, “Unity Gaussian Splatting,” 2023. [Online]. Available: https://github.com/aras-p/UnityGaussianSplatting | - |
| dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/99656 | - |
| dc.description.abstract | 隨著虛擬實境（VR）與沉浸式媒體的發展，如何有效重建具真實感的人物動作場景，成為影像處理與互動內容設計中的重要課題。現有動態人像捕捉系統雖可產出高品質結果，卻多仰賴昂貴設備與複雜後製流程，應用門檻高。近年新興之高斯潑灑（Gaussian Splatting）技術，提供實現動態人像重建的低成本替代方案。
本研究提出一套基於四維高斯潑灑（4D Gaussian Splatting）之人物動作重建與處理流程。針對說書演講、太極拳教學等以人物為主之表演，先進行多視角同步拍攝，並以 Background Matting V2 擷取每影格前景，保留人物、椅子與道具等作為重建目標。為解決 V^3 原始方法於 3D 重建階段可能殘留背景色點雲之問題，本研究設計一套基於背景顏色的高斯點過濾機制，以去除非目標高斯點並保留動作連貫性。最終，重建成果整合至 Unity 並部署於 Meta Quest 3 平台，支援使用者於 VR 環境中自由視角觀看表演內容。 | zh_TW |
| dc.description.abstract | With the growth of virtual reality (VR) and immersive media, reconstructing realistic human-centric motion has become an important challenge. Gaussian Splatting offers a lightweight alternative for dynamic reconstruction.
This study proposes a pipeline based on 4D Gaussian Splatting for scenes such as storytelling or Tai Chi. Multi-view videos are processed with Background Matting V2 to extract foreground elements. To address the issue of background-colored splats in the original V^3 method, we introduce a color-based filtering strategy to enhance foreground clarity and motion consistency. Finally, the reconstructed results are integrated into Unity and deployed on Meta Quest 3, enabling free-viewpoint playback in VR. | en |
| dc.description.provenance | Submitted by admin ntu (admin@lib.ntu.edu.tw) on 2025-09-17T16:16:56Z No. of bitstreams: 0 | en |
| dc.description.provenance | Made available in DSpace on 2025-09-17T16:16:56Z (GMT). No. of bitstreams: 0 | en |
| dc.description.tableofcontents | 誌謝 i
摘要 iii Abstract iv Contents v List of Figures viii List of Tables ix Chapter 1 Introduction 1 1.1 Background and Motivation . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Existing Methods and Challenges . . . . . . . . . . . . . . . . . . . 2 1.3 Research Objectives and Contributions . . . . . . . . . . . . . . . . 2 1.4 Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Chapter 2 Related Work 4 2.1 Gaussian Splatting and Its Extensions . . . . . . . . . . . . . . . . . 4 2.2 Foreground Extraction . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.3 Camera Parameter Estimation . . . . . . . . . . . . . . . . . . . . . 8 Chapter 3 Methods 10 3.1 Overview of the Proposed System Pipeline . . . . . . . . . . . . . . 10 3.2 Camera Calibration and Undistortion . . . . . . . . . . . . . . . . . 12 3.3 Foreground Extraction . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.4 Training of Dynamic Gaussian Splatting . . . . . . . . . . . . . . . . 14 3.4.1 Training Architecture . . . . . . . . . . . . . . . . . . . . . . . . . 14 3.4.2 Background Color Contamination and Filtering Mechanism . . . . . 14 3.4.3 Multi-GPU Segment-Based Training Strategy . . . . . . . . . . . . 17 3.5 Unity Integration and Playback System . . . . . . . . . . . . . . . . 18 3.6 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Chapter 4 Experiments 21 4.1 Experimental Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 4.2 Scaling Strategy for Foreground Extraction . . . . . . . . . . . . . . 22 4.2.1 Experiment Objective and Setup . . . . . . . . . . . . . . . . . . . 22 4.2.2 Result Presentation and Comparison . . . . . . . . . . . . . . . . . 23 4.2.3 Analysis and Discussion . . . . . . . . . . . . . . . . . . . . . . . 25 4.3 Impact of Background Color Choice on Color-Based Filtering . . . . 26 4.3.1 Experiment Objective and Setup . . . . . . . . . . . . . . . . . . . 26 4.3.2 Result Presentation and Comparison . . . . . . . . . . . . . . . . . 27 4.3.3 Analysis and Discussion . . . . . . . . . . . . . . . . . . . . . . . 31 Chapter 5 Applications 33 5.1 Application Scenario Demonstrations . . . . . . . . . . . . . . . . . 33 5.1.1 Storytelling Performance . . . . . . . . . . . . . . . . . . . . . . . 33 5.1.2 Tai Chi Push Hands Demonstration . . . . . . . . . . . . . . . . . . 35 5.1.3 Stylized Character Presentation: Background Effect . . . . . . . . . 36 5.2 Extended Discussion and Application Potential . . . . . . . . . . . . 37 Chapter 6 Conclusion 39 References 42 | - |
| dc.language.iso | en | - |
| dc.subject | 四維高斯潑灑 | zh_TW |
| dc.subject | 人物體積影片 | zh_TW |
| dc.subject | 多視角影片 | zh_TW |
| dc.subject | 動態重建 | zh_TW |
| dc.subject | 虛擬實境 | zh_TW |
| dc.subject | Multi-view Video | en |
| dc.subject | 4D Gaussian Splatting | en |
| dc.subject | Human-centric Volumetric Video | en |
| dc.subject | Virtual Reality | en |
| dc.subject | Dynamic Reconstruction | en |
| dc.title | 基於四維高斯潑灑之人物動作重建與處理 | zh_TW |
| dc.title | Human Motion Reconstruction and Processing Based on 4D Gaussian Splatting | en |
| dc.type | Thesis | - |
| dc.date.schoolyear | 113-2 | - |
| dc.description.degree | 碩士 | - |
| dc.contributor.oralexamcommittee | 歐陽明;王鈺強;林經堯;陳昱吉 | zh_TW |
| dc.contributor.oralexamcommittee | Ming Ouhyoung;Yu-Chiang Wang;Jin-Yao Lin;Yu-Chi Chen | en |
| dc.subject.keyword | 四維高斯潑灑,人物體積影片,多視角影片,動態重建,虛擬實境, | zh_TW |
| dc.subject.keyword | 4D Gaussian Splatting,Human-centric Volumetric Video,Multi-view Video,Dynamic Reconstruction,Virtual Reality, | en |
| dc.relation.page | 44 | - |
| dc.identifier.doi | 10.6342/NTU202503784 | - |
| dc.rights.note | 同意授權(全球公開) | - |
| dc.date.accepted | 2025-08-12 | - |
| dc.contributor.author-college | 電機資訊學院 | - |
| dc.contributor.author-dept | 資訊工程學系 | - |
| dc.date.embargo-lift | 2030-08-05 | - |
| 顯示於系所單位： | 資訊工程學系 | |
文件中的檔案：
| 檔案 | 大小 | 格式 | |
|---|---|---|---|
| ntu-113-2.pdf 此日期後於網路公開 2030-08-05 | 61.3 MB | Adobe PDF |
系統中的文件，除了特別指名其著作權條款之外，均受到著作權保護，並且保留所有的權利。