浏览全部资源
扫码关注微信
收稿日期：2024-08-22，
修回日期：2024-11-20，
网络出版日期：2024-11-27，
移动端阅览
目的
2
三维场景本征分解尝试将场景分解为反射率与照明的乘积，其分解结果可以用于虚拟物体插入、图像材质编辑、重光照等任务，因此受到了广泛的关注与研究。但是，分解规模较大且布局复杂的室内场景是一个高度病态的问题，获得正确的分解结果具有较高的挑战性。
方法
2
本文基于当前最先进的辐射场表示技术——三维高斯泼溅，提出了一种针对室内场景的本征分解算法，大大提高了室内场景本征分解的精度和效率。为了更好地解耦本征属性，本文基于三维高斯泼溅技术设计了一种针对室内场景的本征分解模型，将场景分解为反射率及偏移、照明和残差项，并引入了新的反射率分段稀疏、照明平滑与色度先验约束，以减少分解过程中的歧义，保证分解结果的合理性。同时本文还利用捕获的深度数据增强场景的几何信息，有利于反射率和照明能够更好地解耦，提高合成图像的质量。
结果
2
本文对来自合成数据集Replica的8个场景和真实数据集ScanNet++的5个场景进行了实验，并对分解结果进行可视化；同时本文还测试了新视角下合成图像的PSNR、SSIM、LPIPS指标。结果显示，本文方法不仅可以得到在视觉上更加合理的分解结果，并且合成图像的上述指标在Replica数据集上平均达到了34.6955、0.9654和0.0861，在ScanNet++数据集上平均达到了27.9496、0.8950和0.1444，优于以往的三维场景本征分解算法。
结论
2
与以往的工作相比，本文提出的方法能够快速分解室内场景，并支持在新视角下推理场景属性和合成高保真的图像，具有较高的应用价值。<p>
目的
2
三维场景本征分解尝试将场景分解为反射率与照明的乘积，其分解结果可以用于虚拟物体插入、图像材质编辑、重光照等任务，因此受到了广泛的关注与研究。但是，分解规模较大且布局复杂的室内场景是一个高度病态的问题，获得正确的分解结果具有较高的挑战性。</p><p>
方法
2
本文基于当前最先进的辐射场表示技术——三维高斯泼溅，提出了一种针对室内场景的本征分解算法，大大提高了室内场景本征分解的精度和效率。为了更好地解耦本征属性，本文基于三维高斯泼溅技术设计了一种针对室内场景的本征分解模型，将场景分解为反射率及偏移、照明和残差项，并引入了新的反射率分段稀疏、照明平滑与色度先验约束，以减少分解过程中的歧义，保证分解结果的合理性。同时本文还利用捕获的深度数据增强场景的几何信息，有利于反射率和照明能够更好地解耦，提高合成图像的质量。</p><p>
结果
2
本文对来自合成数据集Replica的8个场景和真实数据集NYU的3个场景进行了实验，并对分解结果进行可视化；同时本文还测试了新视角下合成图像的PSNR、SSIM、LPIPS指标。结果显示，本文方法不仅可以得到在视觉上更加合理的分解结果，并且合成图像的上述指标在Replica数据集上平均达到了34.6955、0.9654和0.0861，在NYU数据集上平均达到了28.0148、0.8651和0.1849，优于以往的三维场景本征分解算法。</p><p>
结论
2
与以往的工作相比，本文提出的方法能够快速分解室内场景，并支持在新视角下推理场景属性和合成高保真的图像，具有较高的应用价值。</p>
Objective
2
Intrinsic decomposition of 3D scenes involves breaking down a scene into the product of reflectance and shading. The results of decomposition can be applied to tasks such as virtual object insertion， image material replacement， and relighting， thus have received widespread attention and research. However， decomposing large-scale and intricately structured indoor scenes presents a highly ill-posed problem due to the complexity of indoor environments， which contain various light sources， materials， and geometries. This complexity makes it difficult to accurately disentangle reflectance and shading because multiple possible combinations of these factors can explain the same observed image. Traditional methods often use prior constraints based on physical knowledge， human intuition， and visual perception to facilitate the reasonable separation of intrinsic attributes. Despite these efforts， traditional methods still struggle to handle the complex interaction between reflectance and shading， resulting in poor decomposition results and limited applicability to different scenarios. Therefore， to achieve accurate and consistent decomposition results， it is essential to address the inherent ambiguities and uncertainties involved.
Method
2
To address these challenges， this paper proposes an intrinsic decomposition algorithm specifically designed for indoor scenes， leveraging the state-of-the-art radiance field representation technique called 3D Gaussian splatting. This approach significantly improves the accuracy and efficiency of intrinsic decomposition for complex indoor environments. At present， there are many studies that have further improved the rendering quality and efficiency of 3D Gaussian splatting and applied it to inverse rendering， dynamic scene modeling， and other tasks. This also inspired us to use 3D Gaussian splatting as a proxy for scene representation in this work. Building on this foundation， we developed an intrinsic decomposition model tailored for indoor scenes. This model is based on Retinex theory and aims to decompose the scene into reflectance， representing the albedo of objects， and shading， representing the interaction between lighting and object geometry. However， decomposing the scene into reflectance and shading is not enough to represent all types of objects in indoor scenes. Since there are specular materials in indoor scenes， we use residual terms to fit the specular reflections. To prevent the loss of texture in the reflectance during the optimization process， we use a reflectance offset to capture these details. The above four components effectively represent most objects in indoor scenes and facilitate the decomposition and optimization of intrinsic scene attributes. To better separate the above attributes， we introduce new reflectance piecewise sparsity， shading smoothness， and chromaticity prior constraints to reduce ambiguity in the decomposition process and ensure the rationality of the results. According to the assumptions of Retinex theory， changes in gradient within a scene are attributed to illumination-independent reflectance and shading. Reflectance refers to the inherent color of the object and is the primary factor responsible for significant changes in appearance. Reflectance exhibits sparsity， as it tends to cluster in the RGB color space， forming distinct groups. On the other hand， shading， which is influenced by the slow variation of light intensity in the scene， is generally assumed to be smooth. Since directly separating reflectance from shading is challenging， we adopt methods from other computer vision fields. We piecewise segment the scene， apply sparsity constraints to each reflectance piece， constrain shading to be smoother in flat areas， and use chromaticity priors to ensure the reflectance's color consistency. These constraints effectively ensure the rationality of the decomposition results. In addition to these novel constraints， we enhanced the geometric information of the scene by incorporating captured depth data， which provides valuable cues about the scene's structure. This improved understanding allows for more accurate decoupling of reflectance and shading， ultimately leading to higher-quality synthesized images.
Result
2
We conducted experiments on 8 scenes from the synthetic dataset Replica and 5 scenes from the real-world dataset ScanNet++. The decomposition results were visualized in the main text， and we evaluated the quality of synthesized images from novel views using common image quality metrics， including Peak Signal-to-Noise Ratio （PSNR）， Structural Similarity Index （SSIM）， and Learned Perceptual Image Patch Similarity （LPIPS）. The results demonstrate that our method not only produces more visually plausible decomposition results but also outperforms previous 3D scene intrinsic decomposition algorithms in terms of quantitative metrics. Specifically， on the Replica dataset， our method achieved average PSNR， SSIM， and LPIPS values of 34.6955， 0.9654， and 0.0861， respectively. On the ScanNet++ dataset， our method achieved average values of 27.9496， 0.8950， and 0.1444. These improvements highlight the effectiveness of our approach in handling complex indoor scenes and producing high-fidelity decomposition results. Our experiments also showcased the method's ability to synthesize realistic images from previously unseen views. This capability is particularly important for applications such as augmented reality and virtual reality. The high quality of the synthesized images further demonstrates our method's robustness in decoupling reflectance and illumination.
Conclusion
2
Compared to previous methods， our approach excels in rapidly and accurately decomposing indoor scenes. Its strong generalization across diverse indoor environments—from synthetic setups to real-world datasets—makes it highly valuable for applications such as virtual object insertion， scene relighting， and material editing. This versatility significantly enhances the practical utility of our method. We believe that this work lays a strong foundation for future research， paving the way for more advanced techniques in the field of intrinsic decomposition and beyond.<p>Objective Intrinsic decomposition of 3D scenes involves breaking down a scene into the product of reflectance and shading. The results of decomposition can be applied to tasks such as virtual object insertion， image material replacement， and relighting， thus have received widespread attention and research. However， decomposing large-scale and intricately structured indoor scenes presents a highly ill-posed problem due to the complexity of indoor environments， which contain various light sources， materials， and geometries. This complexity makes it difficult to accurately disentangle reflectance and shading because multiple possible combinations of these factors can explain the same observed image. Traditional methods often use prior constraints based on physical knowledge， human intuition， and visual perception to facilitate the reasonable separation of intrinsic attributes. Despite these efforts， traditional methods still struggle to handle the complex interaction between reflectance and shading， resulting in poor decomposition results and limited applicability to different scenarios. Therefore， to achieve accurate and consistent decomposition results， it is essential to address the inherent ambiguities and uncertainties involved.</p><p>Method To address these challenges， this paper proposes an intrinsic decomposition algorithm specifically designed for indoor scenes， leveraging the state-of-the-art radiance field representation technique called 3D Gaussian splatting. This approach significantly improves the accuracy and efficiency of intrinsic decomposition for complex indoor environments. 3D Gaussian splatting models the scene using 3D Gaussian distribution functions， effectively capturing the intricate details of a scene's radiance field. This technique allows for a more compact and flexible representation of the scene， enabling more accurate reconstruction of the scene's appearance from various viewpoints. At present， there are many studies that have further improved the rendering quality and efficiency of 3D Gaussian splatting and applied it to inverse rendering， dynamic scene modeling， and other tasks. This also inspired us to use 3D Gaussian splatting as a proxy for scene representation in this work. Building on this foundation， we developed an intrinsic decomposition model tailored for indoor scenes. This model is based on Retinex theory and aims to decompose the scene into reflectance， representing the albedo of objects， and shading， representing the interaction between lighting and object geometry. However， decomposing the scene into reflectance and shading is not enough to represent all types of objects in indoor scenes. Since there are specular materials in indoor scenes， we use residual terms to fit the specular reflections. To prevent the loss of texture in the reflectance during the optimization process， we use a reflectance offset to capture these details. The above four components effectively represent most objects in indoor scenes and facilitate the decomposition and optimization of intrinsic scene attributes. To better separate the above attributes， we introduce new reflectance piecewise sparsity， shading smoothness， and chromaticity prior constraints to reduce ambiguity in the decomposition process and ensure the rationality of the results. According to the assumptions of Retinex theory， changes in gradient within a scene are attributed to illumination-independent reflectance and shading. Reflectance refers to the inherent color of the object and is the primary factor responsible for significant changes in appearance. Reflectance exhibits sparsity， as it tends to cluster in the RGB color space， forming distinct groups. On the other hand， shading， which is influenced by the slow variation of light intensity in the scene， is generally assumed to be smooth. Since directly separating reflectance from shading is challenging， we adopt methods from other computer vision fields. We piecewise segment the scene， apply sparsity constraints to each reflectance piece， constrain shading to be smoother in flat areas， and use chromaticity priors to ensure the reflectance's color consistency. These constraints effectively ensure the rationality of the decomposition results. In addition to these novel constraints， we enhanced the geometric information of the scene by incorporating captured depth data， which provides valuable cues about the scene's structure. This improved understanding allows for more accurate decoupling of reflectance and shading， ultimately leading to higher-quality synthesized images.</p>
Akimoto N ， Zhu H C ， Jin Y H and Aoki Y . 2020 . Fast soft color segmentation // Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . Seattle， USA ： IEEE： 8277 - 8286 ［ DOI： 10.1109/CVPR42600.2020.00830 http://dx.doi.org/10.1109/CVPR42600.2020.00830 ］
Aksoy Y ， Aydin T O ， Smolić A and Pollefeys M . 2017 . Unmixing-based soft color segmentation for image manipulation . ACM Transactions on Graphics ， 36 （ 2 ）： 1 - 19 ［ DOI： 1 0.1145/3002176 http://dx.doi.org/10.1145/3002176 ］
Bell S ， Bala K and Snavely N . 2014 . Intrinsic images in the wild . ACM Transactions on Graphics ， 33 （ 4 ）： 1 - 12 ［ DOI： 10.1145/2601097.2601206 http://dx.doi.org/10.1145/2601097.2601206 ］
Bi S ， Han X G and Yu Y Z . 2015 . An L1 image transform for edge-preserving smoothing and scene-level intrinsic decomposition . ACM Transactions on Graphics ， 34 （ 4 ）： 1 - 12 ［ DOI： 10.1145/2766946 http://dx.doi.org/10.1145/2766946 ］
Chen G K and Wang W G . 2024 . A survey on 3D Gaussian splatting ［EB/OL］. ［ 2024-1-22 ］. https://arxiv.org/pdf/2401.03890 https://arxiv.org/pdf/2401.03890
Chen Q F and Koltun V . 2013 . A simple model for intrinsic image decomposition with depth cues // Proceedings of the IEEE international conference on computer vision . Sydney， Australia ： IEEE： 241 - 248 ［ DOI： 10.1109/ICCV.2013.37 http://dx.doi.org/10.1109/ICCV.2013.37 ］
Eigen D ， Puhrsch C and Fergus R . 2014 . Depth map prediction from a single image using a multi-scale deep network . Advances in neural information processing systems ， 27 . ［ DOI： 10.1117/12.2623204 http://dx.doi.org/10.1117/12.2623204 ］
Fei B ， Xu J Y ， Zhang R ， Zhou Q Y ， Yang W D and He Y . 2024 . 3 D Gaussian as a new era： A Survey ［EB/OL］. ［ 2024-7-10 ］. https://arxiv.org/pdf/2402.07181 https://arxiv.org/pdf/2402.07181
Gao J ， Gu C ， Lin Y T ， Li Z H ， Zhu H ， Cao X ， Zhang L and Yao Y . 2024 . Relightable 3D Gaussians： realistic point cloud relighting with BRDF decomposition and ray tracing ［EB/OL］. ［ 2024-8-8 ］. https://arxiv.org/pdf/2311.16043v2 https://arxiv.org/pdf/2311.16043v2
Garces E ， Munoz A ， Lopez‐Moreno J and Gutierrez D . 2012 . Intrinsic images by clustering . Computer graphics forum ， 31 （ 4 ）： 1415 - 1424 ［ DOI： 10.1111/j.1467-8659.2012.03137.x http://dx.doi.org/10.1111/j.1467-8659.2012.03137.x ］
Gong B C ， Wang Y H ， Han X G and Dou Q . 2023 . RecolorNeRF： layer decomposed radiance fields for efficient color editing of 3D scenes // Proceedings of the 31st ACM International Conference on Multimedia . Ottawa， Canada ： ACM： 8004 – 8015 ［ DOI： 10.1145/3581783.3611957 http://dx.doi.org/10.1145/3581783.3611957 ］
Kerbl B ， Kopanas G ， Leimkühler T and Drettakis G . 2023 . 3D Gaussian splatting for real-time radiance field rendering . ACM Transactions on Graphics ， 42 （ 4 ）： 139 ： 1 - 139 ： 14 ［ DOI： 10.1145/3592433 http://dx.doi.org/10.1145/3592433 ］
Kocsis P ， Sitzmann V and Nießner M . 2024 . Intrinsic image diffusion for indoor single-view material estimation // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Seattle， USA ： IEEE： 5198 - 5208 ［ DOI： 10.1109/CVPR52733.2024.00497 http://dx.doi.org/10.1109/CVPR52733.2024.00497 ］
Kuang Z F ， Luan F J ， Bi S ， Shu Z X ， Wetzstein G and Sunkavalli K . 2023 . PaletteNeRF： palette-based appearance editing of neural radiance fields // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Vancouver， Canada ： IEEE： 20691 - 20700 ［ DOI： 10.1109/CVPR52729.2023.01982 http://dx.doi.org/10.1109/CVPR52729.2023.01982 ］
Land E H and McCann J J . 1971 . Lightness and retinex theory . Josa ， 61 （ 1 ）： 1 - 11 ［ DOI： 10.1364/JOSA.61.000001 http://dx.doi.org/10.1364/JOSA.61.000001 ］
Li Z Q and Snavely N . 2018 . CGIntrinsics： better intrinsic image decomposition through physically-based rendering // Proceedings of the 15th European Conference on Computer Vision . Munich， Germany ： Springer： 381 – 399 . ［ DOI： 10.1007/978-3-030-01219-9_23 http://dx.doi.org/10.1007/978-3-030-01219-9_23 ］
Li Z Q ， Shafiei M ， Ramamoorthi R ， Sunkavalli K and Chandraker M . 2020 . Inverse rendering for complex indoor scenes ： shape， spatially-varying lighting and SVBRDF from a single image// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Seattle， WA， USA ： IEEE： 2472 - 2481 ［ DOI： 10.1109/CVPR42600.2020.00255 http://dx.doi.org/10.1109/CVPR42600.2020.00255 ］
Li Z Q ， Shi J ， Bi S ， Zhu R ， Sunkavalli K ， Hasan M ， Xu Z X ， Ramamoorthi R and Chandraker M ， 2022 . Physically-based editing of indoor scene lighting from a single image // Proceedings of the 17th European Conference on Computer Vision . Tel Aviv， Israel ： Spring： 555 - 572 ［ DOI： 10.1007/978-3-031-20068-7_32 http://dx.doi.org/10.1007/978-3-031-20068-7_32 ］
Jiang Y W Q ， Tu J D ， Liu Y ， Gao X F ， Long X X ， Wang W P and Ma Y X . 2024 . GaussianShader： 3D Gaussian Splatting with Shading Functions for Reflective Surfaces // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Seattle， USA ： IEEE： 5322 - 5332 ［ DOI： 10.1109/cvpr52733.2024.00509 http://dx.doi.org/10.1109/cvpr52733.2024.00509 ］
Liang Z H ， Zhang Q ， Feng Y ， Shan Y and Jia K . 2024 . GS-IR： 3D Gaussian splatting for inverse rendering // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Seattle， USA ： IEEE： 21644 - 21653 ［ DOI： 10.1109/CVPR52733.2024.02045 http://dx.doi.org/10.1109/CVPR52733.2024.02045 ］
Luiten J ， Kopanas G ， Leibe B ， and Ramanan D . 2023 . Dynamic 3D Gaussians： tracking by persistent dynamic view synthesis // 2024 International Conference on 3D Vision . Davos， Switzerland ： IEEE： 800 - 809 ［ DOI： 10.1109/3DV62453.2024.00044 http://dx.doi.org/10.1109/3DV62453.2024.00044 ］
Luo J D ， Huang Z Y ， Li Y J ， Zhou X W ， Zhang G F and Bao H J . 2020 . NIID-Net： adapting surface normal knowledge for intrinsic image decomposition in indoor scenes . IEEE Transactions on Visualization and Computer Graphics . 26 （ 12 ）： 3434 – 45 ［ DOI： 10.1109/TVCG.2020.3023565 http://dx.doi.org/10.1109/TVCG.2020.3023565 ］
Luo J D ， Ceylan D ， Yoon J S ， Zhao N X ， Philip J ， Frühstück A ， Li W B ， Richardt C and Wang T F . 2024 . IntrinsicDiffusion： Joint Intrinsic Layers from Latent Diffusion Models // ACM SIGGRAPH 2024 Conference Papers . Denver， USA ： ACM： 1 - 11 ［ DOI： 10.1145/3641519.3657472 http://dx.doi.org/10.1145/3641519.3657472 ］
Ma L ， Agrawal V ， Turki H ， Turki H ， Kim C ， Gao C ， Sander P ， Zollhöfer M and Richardt C . 2024 . SpecNeRF： Gaussian directional encoding for specular reflections // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Seattle， USA ： IEEE： 21188 - 21198 ［ DOI： 10.1109/CVPR52733.2024.02002 http://dx.doi.org/10.1109/CVPR52733.2024.02002 ］
Meka A ， Zollhöfer M ， Richardt C and Theobalt C . 2016 . Live intrinsic video . ACM Transactions on Graphics ， 35 （ 4 ）： 1 - 14 ［ DOI： 10.1145/2897824.2925907 http://dx.doi.org/10.1145/2897824.2925907 ］
Mildenhall B ， Srinivasan P P ， Tancik M ， Barron J T ， Ramamoorthi R and Ng R . 2020 . NeRF： representing scenes as neural radiance fields for view synthesis. Proceedings of the 16th European Conference on Computer Vision . Munich， Germany ： Springer ： 405 – 421 . ［ DOI： 10.1007/978-3-030-58452-8_24 http://dx.doi.org/10.1007/978-3-030-58452-8_24 ］
Schonberger J L and Frahm J M . 2016 . Structure-from-motion revisited // Proceedings of the IEEE conference on computer vision and pattern recognition . Las Vegas， USA ： IEEE： 4104 - 4113 ［ DOI： 10.1109/CVPR.2016.445 http://dx.doi.org/10.1109/CVPR.2016.445 ］
Shi Y H ， Wu Y M ， Wu C M ， Liu X ， Zhao C ， Feng H C ， Zhang J ， Zhou B ， Ding E R and Wang J D . 2024 . GIR： 3D Gaussian inverse rendering for relightable scene factorization ［EB/OL］. ［ 2024-8-15 ］. https://arxiv.org/pdf/2312.05133 https://arxiv.org/pdf/2312.05133
Straub J ， Whelan T ， Ma L N ， Chen Y F ， Wijmans E ， Green S ， Engel J J ， Mur-Artal R ， Ren C ， Verma S ， Clarkson A ， Yan M F ， Budge B ， Yan Y J ， Pan X Q ， Yon J ， Zou Y Y ， Leon K ， Carter N ， Briales J ， Gillingham T ， Mueggler E ， Pesqueira L ， Savva M ， Batra D ， Strasdat H M ， Nardi R D ， Goesele M ， Lovegrove S and Newcombe R . 2019 . The Replica dataset： a digital replica of indoor spaces ［EB/OL］. ［ 2019-7-13 ］. https://arxiv.org/pdf/1906.05797 https://arxiv.org/pdf/1906.05797
Sun J M ， Wu T ， Yang Y L ， Lai Y K and Gao L . 2023 . SOL-NeRF： Sunlight modeling for outdoor scene decomposition and relighting // SIGGRAPH Asia 2023 Conference Papers . Sydney， Australia ： ACM： 1 - 11 .［ DOI： 10.1145/3610548.3618143 http://dx.doi.org/10.1145/3610548.3618143 ］
Wang Y J ， Fan Q N ， Li K ， Chen D D ， Yang J Y ， Lu J Z ， Dani L and Chen B Q . 2022 . High quality rendered dataset and non-local graph convolutional network for intrinsic image decomposition . Journal of Image and Graphics ， 27 （ 2 ）： 404 - 420
王玉洁 ， 樊庆楠 ， 李坤 ， 陈冬冬 ， 杨敬钰 ， 卢健智 ， Dani Lischinski ， 陈宝权 . 2022 . 面向本征图像分解的高质量渲染数据集与非局部卷积网络 . 中国图象图形学报 ， 27 （ 2 ）： 404 - 420 ［ DOI： 10.11834/jig.210705 http://dx.doi.org/10.11834/jig.210705 ］
Wu L W ， Zhu R ， Yaldiz M B ， Zhu Y H ， Cai H ， Matai J ， Porikli F ， Li T M ， Chandraker M and Ramamoorthi R . 2023 . Factorized inverse path tracing for efficient and accurate material-lighting estimation // Proceedings of the IEEE/CVF International Conference on Computer Vision . Paris， France ： IEEE： 3848 - 3858 ［ DOI： 10.1109/ICCV51070.2023.00356 http://dx.doi.org/10.1109/ICCV51070.2023.00356 ］
Wu T ， Sun J M ， Lai Y K and Gao L . 2023 . DE-NeRF： decoupled neural radiance fields for view-consistent appearance editing and high-frequency environmental relighting // ACM SIGGRAPH 2023 conference proceedings . Los Angeles CA USA ： ACM： 1 - 11 ［ DOI： 10.1145/3588432.3591483 http://dx.doi.org/10.1145/3588432.3591483 ］
Wu T ， Sun J M ， Lai Y K ， Ma Y W ， Kobbelt L and Gao L . 2024 . DeferredGS： decoupled and editable gaussian splatting with deferred shading ［EB/OL］. ［ 2024-5-6 ］. https:// https://arxiv.org/pdf/2404.09412 https://https://arxiv.org/pdf/2404.09412
Wu T ， Yuan Y J ， Zhang L X ， Yang J ， Cao Y P ， Yan L Q and Gao L . 2024 . Recent advances in 3D Gaussian splatting . Computational Visual Media . 10 （ 4 ）： 613 - 642 ［ DOI： 10.1007/s41095-024-0436-y http://dx.doi.org/10.1007/s41095-024-0436-y ］
Xing G Y ， Liu Y L ， Zhang W F ， and Ling H B . 2016 . Light mixture intrinsic image decomposition based on a single RGB-D image . The Visual Computer ， 32 ： 1013 - 1023 ［ DOI： 10.1007/s00371-016-1238-8 http://dx.doi.org/10.1007/s00371-016-1238-8 ］
Xing G Y ， Yuan X ， Li L and Liu Y L . 2017 . User-assisted intrinsic image decomposition based on a single RGB-D image . Journal of Computer-Aided Design & Computer Graphics ， 29 （ 07 ）： 1283 - 1291
邢冠宇 ， 袁霞 ， 李龙 ， 刘艳丽 . 2017 . 计算机辅助设计与图形学学报 ， 29 （ 07 ）： 1283 - 1291 ［ DOI： 10.3969/j.issn.1003-9775.2017.07.015 http://dx.doi.org/10.3969/j.issn.1003-9775.2017.07.015 ］
Xing J K and Xu K . 2024 . Physically based differentiable rendering： a survey . Journal of Image and Graphics ， 29 （ 10 ）： 2912 - 2925
邢健开 ， 徐昆 . 2024 . 基于物理的可微渲染综述 . 中国图象图形学报 ， 29 （ 10 ）： 2912 - 2925 ［ DOI： 10.11834/jig.230715 http://dx.doi.org/10.11834/jig.230715 ］
Xiao Q ， Chen M L ， Zhang Y ， Huang X H . 2024 . Neural radiance field reconstruction for sparse indoor panoramas . Journal of Image and Graphics ， 29 （ 09 ）： 2596 - 2609
肖强 ， 陈铭林 ， 张晔 ， 黄小红 . 2024 . 室内稀疏全景图的神经辐射场重建 . 中国图象图形学报 ， 29 （ 09 ）： 2596 - 2609 ［ DOI： 10.11834/jig.230643 http://dx.doi.org/10.11834/jig.230643 ］
Ye W C ， Chen S ， Bao C ， Bao H J ， Pollefeys M ， Cui Z P and Zhang G F . 2023 . IntrinsicNeRF： learning intrinsic neural radiance fields for editable Novel View synthesis // Proceedings of the IEEE/CVF International Conference on Computer Vision . Paris， France ： IEEE： 339 - 351 ［ DOI： 10.1109/ICCV51070.2023.00038 http://dx.doi.org/10.1109/ICCV51070.2023.00038 ］
Yeshwanth C ， Liu Y C ， Nießner M ， Dai A . 2023 . ScanNet++： a high-fidelity dataset of 3d indoor scenes // Proceedings of the IEEE/CVF International Conference on Computer Vision . Paris， France ： IEEE： 339 - 351 ［ DOI： 10.1109/ICCV51070.2023.00008 http://dx.doi.org/10.1109/ICCV51070.2023.00008 ］
Yao Y ， Zhang J Y ， Liu J B ， Qu Y H ， Fang T ， McKinnon D ， Tsin Y and Quan L . 2022 . NeILF： neural incident light field for physically-based material estimation // Proceedings of the 17th European Conference on Computer Vision . Tel Aviv， Israel ： Springer： 700 - 716 . ［ DOI： 10.1007/978-3-031-19821-2_40 http://dx.doi.org/10.1007/978-3-031-19821-2_40 ］
Zeng Z ， Deschaintre V ， Georgiev I ， Hold-Geoffroy Y ， Hu Y W ， Luan F J ， Yan L Q and Hašan M . 2024 . RGB↔X： Image decomposition and synthesis using material-and lighting-aware diffusion models // ACM SIGGRAPH 2024 Conference Papers . Denver， USA ： ACM： 1 - 11 ［ DOI： 10.1145/3641519.3657445 http://dx.doi.org/10.1145/3641519.3657445 ］
Zhang J Y ， Yao Y ， Li S W ， Liu J B ， Fang T ， McKinnon D ， Tsin Y ， Quan L . 2023 . NeILF++： inter-reflectable light fields for geometry and material estimation // Proceedings of the IEEE/CVF International Conference on Computer Vision . Paris， France . IEEE ： 3601 - 3610 ［ DOI： 10.1109/ICCV51070.2023.00333 http://dx.doi.org/10.1109/ICCV51070.2023.00333 ］
Zwicker M ， Pfister H ， Van Baar J ， and Gross M . 2001 . EWA volume splatting // Proceedings Visualization . 2001. San Diego， CA， USA ： IEEE： 29 - 538 ［ DOI： 10.1109/VISUAL.2001.964490 http://dx.doi.org/10.1109/VISUAL.2001.964490 ］
相关作者
相关机构
京公网安备11010802024621