{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,9]],"date-time":"2026-05-09T00:18:46Z","timestamp":1778285926706,"version":"3.51.4"},"reference-count":41,"publisher":"Institution of Engineering and Technology (IET)","issue":"1","license":[{"start":{"date-parts":[[2025,9,26]],"date-time":"2025-09-26T00:00:00Z","timestamp":1758844800000},"content-version":"vor","delay-in-days":268,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2025,1,1]],"date-time":"2025-01-01T00:00:00Z","timestamp":1735689600000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"content-domain":{"domain":["ietresearch.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["IET Image Processing"],"published-print":{"date-parts":[[2025,1]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                  <jats:p>3D Gaussian splatting (3DGS) is an innovative rendering technique that surpasses the neural radiance field (NeRF) in both rendering speed and visual quality by leveraging an explicit 3D scene representation. Existing 3DGS approaches require a large number of calibrated views to generate a consistent and complete scene representation. When input views are limited, 3DGS tends to overfit the training views, leading to noticeable degradation in rendering quality. To address this limitation, we propose a point\u2010wise feature\u2010aware Gaussian splatting framework that enables real\u2010time, high\u2010quality rendering from sparse training views. Specifically, we employ the latest stereo foundation model to estimate accurate camera poses and reconstruct a dense point cloud for Gaussian initialisation. Then we encode the colour attributes of each 3D Gaussian by sampling and aggregating multiscale 2D appearance features from sparse inputs. To enhance point\u2010wise appearance representation, we design a point interaction network based on a self\u2010attention mechanism, allowing each Gaussian point to interact with its nearest neighbours. These enriched features are subsequently decoded into Gaussian parameters through two lightweight multilayer perceptrons for final rendering. Extensive experiments on diverse benchmarks demonstrate that our method significantly outperforms NeRF\u2010based approaches and achieves competitive performance under few\u2010shot settings compared to the state\u2010of\u2010the\u2010art 3DGS\u00a0methods.<\/jats:p>","DOI":"10.1049\/ipr2.70216","type":"journal-article","created":{"date-parts":[[2025,9,26]],"date-time":"2025-09-26T18:40:56Z","timestamp":1758912056000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["PointGS: Point\u2010Wise Feature\u2010Aware Gaussian Splatting for Sparse View Synthesis"],"prefix":"10.1049","volume":"19","author":[{"given":"Lintao","family":"Xiang","sequence":"first","affiliation":[{"name":"Department of Electrical and Electronic Engineering The University of Manchester Manchester UK"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hongpei","family":"Zheng","sequence":"additional","affiliation":[{"name":"Department of Electrical and Electronic Engineering The University of Manchester Manchester UK"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yating","family":"Huang","sequence":"additional","affiliation":[{"name":"Department of Electrical and Electronic Engineering The University of Manchester Manchester UK"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Qijun","family":"Yang","sequence":"additional","affiliation":[{"name":"Department of Electrical and Electronic Engineering The University of Manchester Manchester UK"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9198-5401","authenticated-orcid":false,"given":"Hujun","family":"Yin","sequence":"additional","affiliation":[{"name":"Department of Electrical and Electronic Engineering The University of Manchester Manchester UK"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"265","published-online":{"date-parts":[[2025,9,26]]},"reference":[{"key":"e_1_2_9_2_1","doi-asserted-by":"publisher","DOI":"10.1145\/3503250"},{"key":"e_1_2_9_3_1","doi-asserted-by":"crossref","unstructured":"J. T.Barron B.Mildenhall M.Tancik P.Hedman R.Martin\u2010Brualla andP. P.Srinivasan \u201cMip\u2010NeRF: A Multiscale Representation for Anti\u2010Aliasing Neural Radiance Fields \u201d inProceedings of the IEEE\/CVF International Conference on Computer Vision(IEEE 2021) 5855\u20135864.","DOI":"10.1109\/ICCV48922.2021.00580"},{"key":"e_1_2_9_4_1","doi-asserted-by":"crossref","unstructured":"J. T.Barron B.Mildenhall D.Verbin P. P.Srinivasan andP.Hedman \u201cMip\u2010NeRF 360: Unbounded Anti\u2010Aliased Neural Radiance Fields \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 5470\u20135479.","DOI":"10.1109\/CVPR52688.2022.00539"},{"key":"e_1_2_9_5_1","doi-asserted-by":"publisher","DOI":"10.1145\/3528223.3530127"},{"key":"e_1_2_9_6_1","doi-asserted-by":"publisher","DOI":"10.1145\/3592433"},{"key":"e_1_2_9_7_1","doi-asserted-by":"crossref","unstructured":"Y.Wei S.Liu Y.Rao W.Zhao J.Lu andJ.Zhou \u201cNerfingMVS: Guided Optimization of Neural Radiance Fields for Indoor Multi\u2010View Stereo \u201d inProceedings of the IEEE\/CVF International Conference on Computer Vision(IEEE 2021) 5610\u20135619.","DOI":"10.1109\/ICCV48922.2021.00556"},{"key":"e_1_2_9_8_1","doi-asserted-by":"crossref","unstructured":"B.Roessle J. T.Barron B.Mildenhall P. P.Srinivasan andM.Nie\u00dfner \u201cDense Depth Priors for Neural Radiance Fields From Sparse Input Views \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 12892\u201312901.","DOI":"10.1109\/CVPR52688.2022.01255"},{"key":"e_1_2_9_9_1","unstructured":"S. F.Bhat R.Birkl D.Wofk P.Wonka andM.M\u00fcller \u201cZoedepth: Zero\u2010Shot Transfer by Combining Relative and Metric Depth \u201dpreprint arXiv February 23 2023 https:\/\/doi.org\/10.48550\/arXiv.2302.12288."},{"key":"e_1_2_9_10_1","doi-asserted-by":"crossref","unstructured":"J. L.SchonbergerandJ.\u2010M.Frahm \u201cStructure\u2010From\u2010Motion Revisited \u201d inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition(IEEE 2016) 4104\u20134113.","DOI":"10.1109\/CVPR.2016.445"},{"key":"e_1_2_9_11_1","doi-asserted-by":"crossref","unstructured":"A.Jain M.Tancik andP.Abbeel \u201cPutting NeRF on a Diet: Semantically Consistent Few\u2010Shot View Synthesis \u201d inProceedings of the IEEE\/CVF International Conference on Computer Vision(IEEE 2021) 5885\u20135894.","DOI":"10.1109\/ICCV48922.2021.00583"},{"key":"e_1_2_9_12_1","doi-asserted-by":"crossref","unstructured":"J.Yang M.Pavone andY.Wang \u201cFreeNeRF: Improving Few\u2010Shot Neural Rendering With Free Frequency Regularization \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2023) 8254\u20138263.","DOI":"10.1109\/CVPR52729.2023.00798"},{"key":"e_1_2_9_13_1","doi-asserted-by":"crossref","unstructured":"J.Chung J.Oh andK. M.Lee \u201cDepth\u2010Regularized Optimization for 3D Gaussian Splatting in Few\u2010Shot Images \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2024) 811\u2013820.","DOI":"10.1109\/CVPRW63382.2024.00086"},{"key":"e_1_2_9_14_1","doi-asserted-by":"crossref","unstructured":"Z.Zhu Z.Fan andY.Jiang \u201cFSGS: Real\u2010Time Few\u2010Shot View Synthesis Using Gaussian Splatting \u201dpreprint arXiv December 1 2023 https:\/\/doi.org\/10.48550\/arXiv.2312.00451.","DOI":"10.1007\/978-3-031-72933-1_9"},{"key":"e_1_2_9_15_1","unstructured":"H.Xiong S.Muttukuru R.Upadhyay P.Chari andA.Kadambi \u201cSparsegs: Real\u2010Time 360\u2218${\\circ}$Sparse View Synthesis Using Gaussian Splatting \u201dpreprint arXiv November 30 2023 https:\/\/doi.org\/10.48550\/arXiv.2312.00206."},{"key":"e_1_2_9_16_1","doi-asserted-by":"crossref","unstructured":"J.Wang M.Chen N.Karaev A.Vedaldi andC.Rupprecht \u201cVGGT: Visual Geometry Grounded Transformer \u201dpreprint arXiv March 17 2025 https:\/\/doi.org\/10.48550\/arXiv.2503.11651.","DOI":"10.1109\/CVPR52734.2025.00499"},{"key":"e_1_2_9_17_1","doi-asserted-by":"crossref","unstructured":"J. T.BarronandB.Mildenhall \u201cMip\u2010NeRF 360: Unbounded Anti\u2010Aliased Neural Radiance Fields \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 5470\u20135479.","DOI":"10.1109\/CVPR52688.2022.00539"},{"key":"e_1_2_9_18_1","doi-asserted-by":"crossref","unstructured":"A.Chen Z.Xu F.Zhao et\u00a0al. \u201cMVSNeRF: Fast Generalizable Radiance Field Reconstruction From Multi\u2010View Stereo \u201d inProceedings of the IEEE\/CVF International Conference on Computer Vision(IEEE 2021) 14124\u201314133.","DOI":"10.1109\/ICCV48922.2021.01386"},{"key":"e_1_2_9_19_1","doi-asserted-by":"crossref","unstructured":"J.Zhang F.Zhan Y.Yu andK.Liu \u201cPose\u2010Free Neural Radiance Fields Via Implicit Pose Regularization \u201d inProceedings of the IEEE\/CVF International Conference on Computer Vision(IEEE 2023) 3534\u20133543.","DOI":"10.1109\/ICCV51070.2023.00327"},{"key":"e_1_2_9_20_1","doi-asserted-by":"crossref","unstructured":"X.Huang Q.Zhang andY.Feng \u201cHDR\u2010NeRF: High Dynamic Range Neural Radiance Fields \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 18398\u201318408.","DOI":"10.1109\/CVPR52688.2022.01785"},{"key":"e_1_2_9_21_1","doi-asserted-by":"crossref","unstructured":"S.Fridovich\u2010Keil A.Yu M.Tancik Q.Chen B.Recht andA.Kanazawa \u201cPlenoxels: Radiance Fields Without Neural Networks \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 5501\u20135510.","DOI":"10.1109\/CVPR52688.2022.00542"},{"key":"e_1_2_9_22_1","doi-asserted-by":"crossref","unstructured":"Z.Fan K.Wang K.Wen andZ.Zhu \u201cLightGaussian: Unbounded 3D Gaussian Compression With 15\u00d7$\\times$Reduction and 200+ fps \u201dpreprint arXiv November 29 2023 https:\/\/doi.org\/10.48550\/arXiv.2311.17245.","DOI":"10.52202\/079017-4447"},{"key":"e_1_2_9_23_1","doi-asserted-by":"crossref","unstructured":"J. C.LeeandD.Rho \u201cCompact 3D Gaussian Representation for Radiance Field \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2024) 21719\u201321728.","DOI":"10.1109\/CVPR52733.2024.02052"},{"key":"e_1_2_9_24_1","doi-asserted-by":"crossref","unstructured":"X.Liu X.Wu andP.Zhang \u201cCompgs: Efficient 3d Scene Representation via Compressed Gaussian Splatting \u201dpreprint arXiv April 15 2024 https:\/\/doi.org\/10.48550\/arXiv.2404.09458.","DOI":"10.1145\/3664647.3681468"},{"key":"e_1_2_9_25_1","doi-asserted-by":"crossref","unstructured":"J.Lin Z.Li andX.Tang \u201cVastgaussian: Vast 3D Gaussians for Large Scene Reconstruction \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2024) 5166\u20135175.","DOI":"10.1109\/CVPR52733.2024.00494"},{"key":"e_1_2_9_26_1","doi-asserted-by":"crossref","unstructured":"Y.LiuandH.Guan \u201cCityGaussian: Real\u2010Time High\u2010Quality Large\u2010Scale Scene Rendering With Gaussians \u201dpreprint arXiv April 1 2024 https:\/\/doi.org\/10.48550\/arXiv.2404.01133.","DOI":"10.1007\/978-3-031-72640-8_15"},{"key":"e_1_2_9_27_1","unstructured":"J.Guo X.Ma andY.Fan \u201cSemantic Gaussians: Open\u2010Vocabulary Scene Understanding With 3D Gaussian Splatting \u201dpreprint arXiv March 22 2024 https:\/\/doi.org\/10.48550\/arXiv.2403.15624."},{"key":"e_1_2_9_28_1","doi-asserted-by":"crossref","unstructured":"M.Niemeyer J. T.Barron andB.Mildenhall \u201cRegnerf: Regularizing Neural Radiance Fields for View Synthesis From Sparse Inputs \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 5480\u20135490.","DOI":"10.1109\/CVPR52688.2022.00540"},{"key":"e_1_2_9_29_1","doi-asserted-by":"crossref","unstructured":"K.Deng A.Liu J.\u2010Y.Zhu andD.Ramanan \u201cDepth\u2010Supervised Nerf: Fewer Views and Faster Training for Free \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 12882\u201312891.","DOI":"10.1109\/CVPR52688.2022.01254"},{"key":"e_1_2_9_30_1","doi-asserted-by":"crossref","unstructured":"G.Wang Z.Chen C. C.Loy andZ.Liu \u201cSparsenerf: Distilling Depth Ranking for Few\u2010Shot Novel View Synthesis \u201d inProceedings of the IEEE\/CVF International Conference on Computer Vision(IEEE 2023) 9065\u20139076.","DOI":"10.1109\/ICCV51070.2023.00832"},{"key":"e_1_2_9_31_1","doi-asserted-by":"crossref","unstructured":"J.LiandJ.Zhang \u201cDNGaussian: Optimizing Sparse\u2010View 3D Gaussian Radiance Fields With Global\u2010Local Depth Normalization \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2024) 20775\u201320785.","DOI":"10.1109\/CVPR52733.2024.01963"},{"key":"e_1_2_9_32_1","doi-asserted-by":"crossref","unstructured":"A.Paliwal W.Ye andJ.Xiong \u201cCoherentgs: Sparse Novel View Synthesis With Coherent 3D Gaussians \u201d inEuropean Conference on Computer Vision(Springer 2024) 19\u201337.","DOI":"10.1007\/978-3-031-73404-5_2"},{"key":"e_1_2_9_33_1","doi-asserted-by":"crossref","unstructured":"H.Xu A.Chen Y.Chen et\u00a0al. \u201cMuRF: Multi\u2010Baseline Radiance Fields \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2024) 20041\u201320050.","DOI":"10.1109\/CVPR52733.2024.01894"},{"key":"e_1_2_9_34_1","doi-asserted-by":"crossref","unstructured":"D.Charatan S. L.Li A.Tagliasacchi andV.Sitzmann \u201cPixelsplat: 3D Gaussian Splats From Image Pairs for Scalable Generalizable 3D Reconstruction \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2024) 19457\u201319467.","DOI":"10.1109\/CVPR52733.2024.01840"},{"key":"e_1_2_9_35_1","doi-asserted-by":"crossref","unstructured":"Y.Chen H.Xu C.Zheng et\u00a0al. \u201cMVSPLAT: Efficient 3D Gaussian Splatting From Sparse Multi\u2010View Images \u201d inEuropean Conference on Computer Vision(Springer 2024) 370\u2013386.","DOI":"10.1007\/978-3-031-72664-4_21"},{"key":"e_1_2_9_36_1","doi-asserted-by":"crossref","unstructured":"D.RobertandH.Raguet \u201cEfficient 3D Semantic Segmentation With Superpoint Transformer \u201d inProceedings of the IEEE\/CVF International Conference on Computer Vision(IEEE 2023) 17195\u201317204.","DOI":"10.1109\/ICCV51070.2023.01577"},{"key":"e_1_2_9_37_1","doi-asserted-by":"crossref","unstructured":"J.Park S.Lee S.Kim Y.Xiong andH. J.Kim \u201cSelf\u2010Positioning Point\u2010Based Transformer for Point Cloud Understanding \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2023) 21814\u201321823.","DOI":"10.1109\/CVPR52729.2023.02089"},{"key":"e_1_2_9_38_1","doi-asserted-by":"publisher","DOI":"10.1145\/3306346.3322980"},{"key":"e_1_2_9_39_1","doi-asserted-by":"publisher","DOI":"10.1145\/3272127.3275084"},{"key":"e_1_2_9_40_1","doi-asserted-by":"crossref","unstructured":"R.ZhangandP.Isola \u201cThe Unreasonable Effectiveness of Deep Features as a Perceptual Metric \u201d inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition(IEEE 2018) 586\u2013595.","DOI":"10.1109\/CVPR.2018.00068"},{"key":"e_1_2_9_41_1","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2003.819861"},{"key":"e_1_2_9_42_1","series-title":"Lecture Notes in Computer Science","first-page":"203","volume-title":"Computer Vision \u2010 ECCV 2024","author":"Xu W.","year":"2024"}],"container-title":["IET Image Processing"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/pdf\/10.1049\/ipr2.70216","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/full-xml\/10.1049\/ipr2.70216","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/pdf\/10.1049\/ipr2.70216","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,9]],"date-time":"2026-05-09T00:06:21Z","timestamp":1778285181000},"score":1,"resource":{"primary":{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/10.1049\/ipr2.70216"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,1]]},"references-count":41,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2025,1]]}},"alternative-id":["10.1049\/ipr2.70216"],"URL":"https:\/\/doi.org\/10.1049\/ipr2.70216","archive":["Portico"],"relation":{},"ISSN":["1751-9659","1751-9667"],"issn-type":[{"value":"1751-9659","type":"print"},{"value":"1751-9667","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,1]]},"assertion":[{"value":"2025-06-16","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-09-13","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-09-26","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70216"}}