{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,21]],"date-time":"2026-05-21T11:06:01Z","timestamp":1779361561170,"version":"3.51.4"},"reference-count":56,"publisher":"Wiley","issue":"7","license":[{"start":{"date-parts":[[2024,11,7]],"date-time":"2024-11-07T00:00:00Z","timestamp":1730937600000},"content-version":"vor","delay-in-days":37,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62072020"],"award-info":[{"award-number":["62072020"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100019612","name":"Project of Qingdao Leading Talents in Entrepreneurship and Innovation","doi-asserted-by":"publisher","award":["19\u20103\u20102\u201021\u2010zhc"],"award-info":[{"award-number":["19\u20103\u20102\u201021\u2010zhc"]}],"id":[{"id":"10.13039\/501100019612","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62201018"],"award-info":[{"award-number":["62201018"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Computer Graphics Forum"],"published-print":{"date-parts":[[2024,10]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:p>This paper introduces a novel continual learning framework for synthesising novel views of multiple scenes, learning multiple 3D scenes incrementally, and updating the network parameters only with the training data of the upcoming new scene. We build on Neural Radiance Fields (NeRF), which uses multi\u2010layer perceptron to model the density and radiance field of a scene as the implicit function. While NeRF and its extensions have shown a powerful capability of rendering photo\u2010realistic novel views in a single 3D scene, managing these growing 3D NeRF assets efficiently is a new scientific problem. Very few works focus on the efficient representation or continuous learning capability of multiple scenes, which is crucial for the practical applications of NeRF. To achieve these goals, our key idea is to represent multiple scenes as the linear combination of a cross\u2010scene weight matrix and a set of scene\u2010specific weight matrices generated from a global parameter generator. Furthermore, we propose an uncertain surface knowledge distillation strategy to transfer the radiance field knowledge of previous scenes to the new model. Representing multiple 3D scenes with such weight matrices significantly reduces memory requirements. At the same time, the uncertain surface distillation strategy greatly overcomes the catastrophic forgetting problem and maintains the photo\u2010realistic rendering quality of previous scenes. Experiments show that the proposed approach achieves state\u2010of\u2010the\u2010art rendering quality of continual learning NeRF on NeRF\u2010Synthetic, LLFF, and TanksAndTemples datasets while preserving extra low storage cost.<\/jats:p>","DOI":"10.1111\/cgf.15255","type":"journal-article","created":{"date-parts":[[2024,11,8]],"date-time":"2024-11-08T07:03:24Z","timestamp":1731049404000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":15,"title":["SCARF: Scalable Continual Learning Framework for Memory\u2010efficient Multiple Neural Radiance Fields"],"prefix":"10.1111","volume":"43","author":[{"ORCID":"https:\/\/orcid.org\/0009-0000-7676-3408","authenticated-orcid":false,"given":"Yuze","family":"Wang","sequence":"first","affiliation":[{"name":"State Key Laboratory of Virtual Reality Technology and Systems, School of Computer Science and Engineering Beihang University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3191-1662","authenticated-orcid":false,"given":"Junyi","family":"Wang","sequence":"additional","affiliation":[{"name":"State Key Laboratory of Virtual Reality Technology and Systems, School of Computer Science and Engineering Beihang University"},{"name":"School of Computer Science and Technology Shandong University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4334-6103","authenticated-orcid":false,"given":"Chen","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Computer and Artificial Intelligence Beijing Technology and Business University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-8238-0670","authenticated-orcid":false,"given":"Wantong","family":"Duan","sequence":"additional","affiliation":[{"name":"State Key Laboratory of Virtual Reality Technology and Systems, School of Computer Science and Engineering Beihang University"},{"name":"Jingdezhen Research Institute of Beihang University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1010-7229","authenticated-orcid":false,"given":"Yongtang","family":"Bao","sequence":"additional","affiliation":[{"name":"Shandong University of Science and Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9304-1933","authenticated-orcid":false,"given":"Yue","family":"Qi","sequence":"additional","affiliation":[{"name":"State Key Laboratory of Virtual Reality Technology and Systems, School of Computer Science and Engineering Beihang University"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"311","published-online":{"date-parts":[[2024,11,7]]},"reference":[{"key":"e_1_2_7_2_2","unstructured":"BaoY. DingT. HuoJ. LiW. LiY. GaoY.: Insertnerf: Instilling generalizability into nerf with hypernet modules.arXiv preprint arXiv:2308.13897(2023). 2"},{"key":"e_1_2_7_3_2","first-page":"2372","volume-title":"International Conference on Artificial Intelligence and Statistics","author":"Benzing F.","year":"2022"},{"key":"e_1_2_7_4_2","doi-asserted-by":"crossref","unstructured":"BarronJ. T. MildenhallB. VerbinD. SrinivasanP. P. HedmanP.: Mip-nerf 360: Unbounded anti-aliased neural radiance fields. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2022) pp.5470\u20135479. 1","DOI":"10.1109\/CVPR52688.2022.00539"},{"key":"e_1_2_7_5_2","doi-asserted-by":"crossref","unstructured":"BaoC. ZhangY. YangB. FanT. YangZ. BaoH. ZhangG. CuiZ.: Sine: Semantic-driven image-based nerf editing with prior-guided editing field. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2023) pp.20919\u201320929. 3","DOI":"10.1109\/CVPR52729.2023.02004"},{"key":"e_1_2_7_6_2","first-page":"1240","volume-title":"International conference on machine learning","author":"Caccia L.","year":"2020"},{"key":"e_1_2_7_7_2","doi-asserted-by":"crossref","unstructured":"ChenY. ChenX. WangX. ZhangQ. GuoY. ShanY. WangF.: Local-to-global registration for bundle-adjusting neural radiance fields. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2023) pp.8264\u20138273. 3","DOI":"10.1109\/CVPR52729.2023.00799"},{"key":"e_1_2_7_8_2","doi-asserted-by":"crossref","unstructured":"ChenZ. FunkhouserT. HedmanP. TagliasacchiA.: Mobilenerf: Exploiting the polygon rasterization pipeline for efficient neural field rendering on mobile architectures. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2023) pp.16569\u201316578. 3","DOI":"10.1109\/CVPR52729.2023.01590"},{"key":"e_1_2_7_9_2","unstructured":"ChungJ. LeeK. BaikS. LeeK. M.:Meil-nerf: Memory-efficient incremental learning of neural radiance fields 2022. arXiv:2212.08328. 2 3 7 9"},{"key":"e_1_2_7_10_2","first-page":"333","volume-title":"European Conference on Computer Vision","author":"Chen A.","year":"2022"},{"key":"e_1_2_7_11_2","unstructured":"ChenA. XuZ. ZhaoF. ZhangX. XiangF. YuJ. SuH.: Mvsnerf: Fast generalizable radiance field reconstruction from multi-view stereo. InProceedings of the IEEE\/CVF International Conference on Computer Vision(2021) pp.14124\u201314133. 2 3"},{"key":"e_1_2_7_12_2","doi-asserted-by":"crossref","unstructured":"ChenJ. YiW. MaL. JiaX. LuH.: Gm-nerf: Learning generalizable model-based neural radiance fields from multi-view images. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2023) pp.20648\u201320658. 3","DOI":"10.1109\/CVPR52729.2023.01978"},{"key":"e_1_2_7_13_2","doi-asserted-by":"crossref","unstructured":"DengT. ShenG. QinT. WangJ. ZhaoW. WangJ. WangD. ChenW.: Plgslam: Progressive neural scene represenation with local to global bundle adjustment. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2024) pp.19657\u201319666. 3","DOI":"10.1109\/CVPR52733.2024.01859"},{"key":"e_1_2_7_14_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58621-8_23"},{"key":"e_1_2_7_15_2","first-page":"1126","volume-title":"International conference on machine learning","author":"Finn C.","year":"2017"},{"key":"e_1_2_7_16_2","doi-asserted-by":"publisher","DOI":"10.1145\/3550454.3555458"},{"key":"e_1_2_7_17_2","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v37i1.25135"},{"key":"e_1_2_7_18_2","doi-asserted-by":"crossref","unstructured":"GordonC. ChngS.-F. MacDonaldL. LuceyS.: On quantizing implicit neural representations. InProceedings of the IEEE\/CVF Winter Conference on Applications of Computer Vision(2023) pp.341\u2013350. 3","DOI":"10.1109\/WACV56688.2023.00042"},{"key":"e_1_2_7_19_2","article-title":"Generative adversarial nets","volume":"27","author":"Goodfellow I.","year":"2014","journal-title":"Advances in neural information processing systems"},{"key":"e_1_2_7_20_2","unstructured":"HaD. DaiA. M. LeQ. V.: Hypernetworks.CoRR abs\/1609.09106(2016). URL:http:\/\/arxiv.org\/abs\/1609.09106 arXiv:1609.09106. 3 4 5"},{"key":"e_1_2_7_21_2","first-page":"e14940","volume-title":"Computer Graphics Forum","author":"He H.","year":"2023"},{"key":"e_1_2_7_22_2","unstructured":"HuE. J. ShenY. WallisP. Allen-ZhuZ. LiY. WangS. WangL. ChenW.: Lora: Low-rank adaptation of large language models.arXiv preprint arXiv:2106.09685(2021). 3"},{"key":"e_1_2_7_23_2","doi-asserted-by":"crossref","unstructured":"HeK. ZhangX. RenS. SunJ.: Deep residual learning for image recognition. InProceedings of the IEEE conference on computer vision and pattern recognition(2016) pp.770\u2013778. 2","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_2_7_24_2","unstructured":"KingmaD. P. BaJ.: Adam: A method for stochastic optimization.arXiv preprint arXiv:1412.6980(2014). 8"},{"key":"e_1_2_7_25_2","unstructured":"KendallA. GalY. CipollaR.: Multi-task learning using uncertainty to weigh losses for scene geometry and semantics.CoRR abs\/1705.07115(2017). URL:http:\/\/arxiv.org\/abs\/1705.07115 arXiv:1705.07115. 5"},{"key":"e_1_2_7_26_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1611835114"},{"key":"e_1_2_7_27_2","doi-asserted-by":"publisher","DOI":"10.1145\/3072959.3073599"},{"key":"e_1_2_7_28_2","unstructured":"LindellD. B. MartelJ. N. WetzsteinG.: Autoint: Automatic integration for fast neural volume rendering. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2021) pp.14556\u201314565. 3"},{"key":"e_1_2_7_29_2","doi-asserted-by":"crossref","unstructured":"LiL. ShenZ. WangZ. ShenL. BoL.: Compressing volumetric radiance fields to 1 mb. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2023) pp.4222\u20134231. 9","DOI":"10.1109\/CVPR52729.2023.00411"},{"key":"e_1_2_7_30_2","unstructured":"LiuX. WuC. MentaM. HerranzL. RaducanuB. BagdanovA. D. JuiS. de WeijerJ. v.: Generative feature replay for class-incremental learning. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition Workshops(2020) pp.226\u2013227. 3"},{"key":"e_1_2_7_31_2","first-page":"351","volume-title":"European Conference on Computer Vision","author":"Lin H.","year":"2022"},{"key":"e_1_2_7_32_2","doi-asserted-by":"crossref","unstructured":"Martin-BruallaR. RadwanN. SajjadiM. S. BarronJ. T. DosovitskiyA. DuckworthD.: Nerf in the wild: Neural radiance fields for unconstrained photo collections. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2021) pp.7210\u20137219. 9","DOI":"10.1109\/CVPR46437.2021.00713"},{"key":"e_1_2_7_33_2","doi-asserted-by":"crossref","unstructured":"MallyaA. DavisD. LazebnikS.: Piggyback: Adapting a single network to multiple tasks by learning to mask weights. InProceedings of the European conference on computer vision (ECCV)(2018) pp.67\u201382. 3","DOI":"10.1007\/978-3-030-01225-0_5"},{"key":"e_1_2_7_34_2","doi-asserted-by":"publisher","DOI":"10.1145\/3528223.3530127"},{"key":"e_1_2_7_35_2","doi-asserted-by":"crossref","unstructured":"MallyaA. LazebnikS.: Packnet: Adding multiple tasks to a single network by iterative pruning. InProceedings of the IEEE conference on Computer Vision and Pattern Recognition(2018) pp.7765\u20137773. 7 9","DOI":"10.1109\/CVPR.2018.00810"},{"key":"e_1_2_7_36_2","unstructured":"MishraA. LatorreJ. A. PoolJ. StosicD. StosicD. VenkateshG. YuC. MicikeviciusP.:Accelerating sparse deep neural networks 2021. arXiv:2104.08378. 8"},{"key":"e_1_2_7_37_2","doi-asserted-by":"publisher","DOI":"10.1145\/3306346.3322980"},{"key":"e_1_2_7_38_2","doi-asserted-by":"publisher","DOI":"10.1145\/3503250"},{"key":"e_1_2_7_39_2","unstructured":"PoR. DongZ. BergmanA. W. WetzsteinG.: Instant continual learning of neural radiance fields. InProceedings of the IEEE\/CVF International Conference on Computer Vision(2023) pp.3334\u20133344. 2 3"},{"key":"e_1_2_7_40_2","unstructured":"QuH. RahmaniH. XuL. WilliamsB. LiuJ.: Recent advances of continual learning in computer vision: An overview.arXiv preprint arXiv:2109.11369(2021). 3"},{"key":"e_1_2_7_41_2","unstructured":"RebuffiS.-A. KolesnikovA. SperlG. LampertC. H.: icarl: Incremental classifier and representation learning. InProceedings of the IEEE conference on Computer Vision and Pattern Recognition(2017) pp.2001\u20132010. 3"},{"key":"e_1_2_7_42_2","unstructured":"RusuA. A. RabinowitzN. C. DesjardinsG. SoyerH. KirkpatrickJ. KavukcuogluK. PascanuR. HadsellR.: Progressive neural networks.arXiv preprint arXiv:1606.04671(2016). 3"},{"key":"e_1_2_7_43_2","first-page":"10136","article-title":"Metasdf: Meta-learning signed distance functions","volume":"33","author":"Sitzmann V.","year":"2020","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_7_44_2","unstructured":"SucarE. LiuS. OrtizJ. DavisonA. J.: imap: Implicit mapping and positioning in real-time. InProceedings of the IEEE\/CVF International Conference on Computer Vision(2021) pp.6229\u20136238. 3"},{"key":"e_1_2_7_45_2","unstructured":"ShinS. ParkJ.: Binary radiance fields.arXiv preprint arXiv:2306.07581(2023). 3"},{"key":"e_1_2_7_46_2","doi-asserted-by":"crossref","unstructured":"SunC. SunM. ChenH.-T.: Direct voxel grid optimization: Super-fast convergence for radiance fields reconstruction. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2022) pp.5459\u20135469. 1 9","DOI":"10.1109\/CVPR52688.2022.00538"},{"key":"e_1_2_7_47_2","first-page":"14798","article-title":"Compressible-composable nerf via rank-residual decomposition","volume":"35","author":"Tang J.","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_7_48_2","doi-asserted-by":"crossref","unstructured":"TancikM. MildenhallB. WangT. SchmidtD. SrinivasanP. P. BarronJ. T. NgR.: Learned initializations for optimizing coordinate-based neural representations. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2021) pp.2846\u20132855. 3","DOI":"10.1109\/CVPR46437.2021.00287"},{"key":"e_1_2_7_49_2","article-title":"Attention is all you need","volume":"30","author":"Vaswani A.","year":"2017","journal-title":"Advances in neural information processing systems"},{"key":"e_1_2_7_50_2","doi-asserted-by":"crossref","unstructured":"WanZ. RichardtC. Bo\u017ei\u010dA. LiC. RengarajanV. NamS. XiangX. LiT. ZhuB. RanjanR. et al.: Learning neural duplex radiance fields for real-time view synthesis. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2023) pp.8307\u20138316. 3","DOI":"10.1109\/CVPR52729.2023.00803"},{"key":"e_1_2_7_51_2","doi-asserted-by":"crossref","unstructured":"WangQ. WangZ. GenovaK. SrinivasanP. P. ZhouH. BarronJ. T. Martin-BruallaR. SnavelyN. FunkhouserT.: Ibrnet: Learning multi-view image-based rendering. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2021) pp.4690\u20134699. 2 3","DOI":"10.1109\/CVPR46437.2021.00466"},{"key":"e_1_2_7_52_2","doi-asserted-by":"crossref","unstructured":"WangY. WangJ. QuY. QiY.: Rip-nerf: Learning rotation-invariant point-based neural radiance field for fine-grained editing and compositing. InProceedings of the 2023 ACM International Conference on Multimedia Retrieval(2023) pp.125\u2013134. 3","DOI":"10.1145\/3591106.3592276"},{"key":"e_1_2_7_53_2","unstructured":"WangL. ZhangX. SuH. ZhuJ.: A comprehensive survey of continual learning: Theory method and application.arXiv preprint arXiv:2302.00487(2023). 3"},{"key":"e_1_2_7_54_2","doi-asserted-by":"crossref","unstructured":"YangH. HongL. LiA. HuT. LiZ. LeeG. H. WangL.: Contranerf: Generalizable neural radiance fields for synthetic-to-real novel view synthesis via contrastive learning. InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(2023) pp.16508\u201316517. 2 3","DOI":"10.1109\/CVPR52729.2023.01584"},{"key":"e_1_2_7_55_2","unstructured":"ZhipengCai M. M.: Clnerf: Continual learning meets nerf. InICCV(2023). 2 3 9"},{"key":"e_1_2_7_56_2","unstructured":"ZhangL. LiM. ChenC. XuJ.: Il-nerf: Incremental learning for neural radiance fields with camera pose alignment.arXiv preprint arXiv:2312.05748(2023). 2"},{"key":"e_1_2_7_57_2","first-page":"7693","volume-title":"International Conference on Machine Learning","author":"Zintgraf L.","year":"2019"}],"container-title":["Computer Graphics Forum"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/cgf.15255","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,11,14]],"date-time":"2024-11-14T08:06:22Z","timestamp":1731571582000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1111\/cgf.15255"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,10]]},"references-count":56,"journal-issue":{"issue":"7","published-print":{"date-parts":[[2024,10]]}},"alternative-id":["10.1111\/cgf.15255"],"URL":"https:\/\/doi.org\/10.1111\/cgf.15255","archive":["Portico"],"relation":{},"ISSN":["0167-7055","1467-8659"],"issn-type":[{"value":"0167-7055","type":"print"},{"value":"1467-8659","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,10]]},"assertion":[{"value":"2024-11-07","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e15255"}}