{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,11]],"date-time":"2026-02-11T13:55:28Z","timestamp":1770818128794,"version":"3.50.1"},"reference-count":29,"publisher":"MDPI AG","issue":"2","license":[{"start":{"date-parts":[[2026,2,11]],"date-time":"2026-02-11T00:00:00Z","timestamp":1770768000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Excellent Research and Innovation Teams Project of Universities in Anhui Province","award":["2024AH010030"],"award-info":[{"award-number":["2024AH010030"]}]},{"name":"Key Scientific Research Project of Universities in Anhui Province","award":["2025AHGXZK30878"],"award-info":[{"award-number":["2025AHGXZK30878"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Information"],"abstract":"<jats:p>To address fragile feature representation in sparse regions and detail loss in occluded scenes caused by uneven sampling density in 3D point cloud semantic segmentation on the SemanticKITTI dataset, this article proposes an innovative framework that integrates density-adaptive feature enhancement with lightweight spectral fine-tuning, which involves frequency-domain transformations (e.g., Fast Fourier Transform) applied to point cloud features to optimize computational efficiency and enhance robustness in sparse regions, which involves frequency-domain transformations to optimize features efficiently. The method begins by accurately calculating each point\u2019s local neighborhood density using KD tree radius search, subsequently injecting this as an additional feature channel to enable the network\u2019s adaptation to density variations. A density-aware loss function is then employed, dynamically adjusting the classification loss weights\u2014by approximately 40% in low-density areas\u2014to strongly penalize misclassifications and enhance feature robustness from sparse points. Additionally, a multi-view projection fusion mechanism is introduced that projects point clouds onto multiple 2D views, capturing detailed information via mature 2D models, with the primary focus on semantic segmentation tasks using the SemanticKITTI dataset to ensure task specificity. This information is then fused with the original 3D features through backprojection, thereby complementing geometric relationships and texture details to effectively alleviate occlusion artifacts. Experiments on the SemanticKITTI dataset for semantic segmentation show significant performance improvements over the baseline, achieving Precision 0.91, Recall 0.89, and F1-Score 0.90. In low-density regions, the F1-Score improved from 0.73 to 0.80. Ablation studies highlight the contributions of density feature injection, multi-view fusion, and density-aware loss, enhancing F1-Score by 3.8%, 2.5%, and 5.0%, respectively. This framework offers an effective approach for accurate and robust point cloud analysis through optimized density techniques and spectral domain fine-tuning.<\/jats:p>","DOI":"10.3390\/info17020184","type":"journal-article","created":{"date-parts":[[2026,2,11]],"date-time":"2026-02-11T11:13:15Z","timestamp":1770808395000},"page":"184","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Research on Density-Adaptive Feature Enhancement and Lightweight Spectral Fine-Tuning Algorithm for 3D Point Cloud Analysis"],"prefix":"10.3390","volume":"17","author":[{"ORCID":"https:\/\/orcid.org\/0009-0001-8434-3611","authenticated-orcid":false,"given":"Wenquan","family":"Huang","sequence":"first","affiliation":[{"name":"School of Artificial Intelligence, Anhui University, Hefei 230601, China"},{"name":"School of Intelligent Manufacturing, Anhui Wenda University of Information Engineering, Hefei 231201, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Teng","family":"Li","sequence":"additional","affiliation":[{"name":"School of Artificial Intelligence, Anhui University, Hefei 230601, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Qing","family":"Cheng","sequence":"additional","affiliation":[{"name":"School of Artificial Intelligence, Anhui University, Hefei 230601, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ping","family":"Qi","sequence":"additional","affiliation":[{"name":"School of Artificial Intelligence, Tongling University, Tongling 244061, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jing","family":"Zhu","sequence":"additional","affiliation":[{"name":"College of Art and Design, Nanning University, Nanning 530200, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2026,2,11]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","unstructured":"Chang, C.H., and Kehtarnavaz, N. (2018). A computationally efficient pipeline for 3D point cloud reconstruction from video sequences. Real-Time Image and Video Processing 2018, SPIE.","DOI":"10.1117\/12.2302674"},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"5432","DOI":"10.1109\/LRA.2020.3007440","article-title":"3D-MiniNet: Learning a 2D Representation From Point Clouds for Fast and Efficient 3D LIDAR Semantic Segmentation","volume":"5","author":"Alonso","year":"2020","journal-title":"IEEE Robot. Autom. Lett."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"112915","DOI":"10.1016\/j.engappai.2025.112915","article-title":"An automated framework for converting point cloud data to building information modeling with segmentation and refinement","volume":"163","author":"Chen","year":"2026","journal-title":"Eng. Appl. Artif. Intell."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"106612","DOI":"10.1016\/j.autcon.2025.106612","article-title":"Semantic instance segmentation and automated 3D BIM reconstruction for viaduct using LiDAR point clouds and weakly-supervised learning","volume":"181","author":"Qiao","year":"2026","journal-title":"Autom. Constr."},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Hu, N., Chen, Z., Ma, R., and Liu, H. (2023). Research on the Detection and Measurement of Roughness of Dam Concrete Layers Using 3D Laser Scanning Technology. Sustainability, 15.","DOI":"10.3390\/su15032649"},{"key":"ref_6","doi-asserted-by":"crossref","unstructured":"Zhang, H., and Wang, T. (2022). An efficient method for producing deep learning point cloud datasets based on BIM 3D model and computer simulation. Second International Symposium on Computer Technology and Information Science (ISCTIS 2022), SPIE.","DOI":"10.1117\/12.2653608"},{"key":"ref_7","first-page":"102","article-title":"Obtaining growth information of field peanuts based on 3D LiDAR perception","volume":"44","author":"Hu","year":"2025","journal-title":"J. Huazhong Agric. Univ."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"270","DOI":"10.1016\/j.proeng.2017.05.181","article-title":"Calculation the Spacing of Discontinuities from 3D Point Clouds","volume":"191","author":"Buyer","year":"2017","journal-title":"Procedia Eng."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"8287","DOI":"10.1007\/s00371-023-03237-7","article-title":"Deep learning based computer vision under the prism of 3D point clouds: A systematic review","volume":"40","author":"Tychola","year":"2024","journal-title":"Vis. Comput."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"212206","DOI":"10.1007\/s11432-019-9932-3","article-title":"Accelerating DNN-based 3D point cloud processing for mobile computing","volume":"62","author":"Liu","year":"2019","journal-title":"Sci. China Inf. Sci."},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"95","DOI":"10.1016\/j.autcon.2013.08.007","article-title":"Computing 3D blind spots of construction equipment: Implementation and evaluation of an automated measurement and visualization method utilizing range point cloud data","volume":"36","author":"Ray","year":"2013","journal-title":"Autom. Constr."},{"key":"ref_12","unstructured":"Shen, Y. (2024). Research on 3D Point Cloud Motion Estimation and Prediction Algorithms in Complex Scenes. [Ph.D. Thesis, Nanjing University of Science and Technology]."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"13","DOI":"10.1016\/j.patrec.2015.05.007","article-title":"Computing refined skeletal features from medial point clouds","volume":"76","author":"Kustra","year":"2016","journal-title":"Pattern Recognit. Lett."},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"112528","DOI":"10.1016\/j.patcog.2025.112528","article-title":"Domain-adaptive point cloud semantic segmentation via knowledge-augmented deep learning","volume":"172","author":"Xu","year":"2026","journal-title":"Pattern Recognit."},{"key":"ref_15","unstructured":"Wei, X. (2025). Research on 3D Point Cloud Registration Method Based on Local Feature Extraction. [Master\u2019s Thesis, Hunan University of Technology]."},{"key":"ref_16","unstructured":"Peng, S. (2025). Design and Implementation of Collaborative SLAM System Based on Deep Point Cloud Compression Technology. [Master\u2019s Thesis, University of Electronic Science and Technology of China]."},{"key":"ref_17","unstructured":"Du, Z. (2025). Research on Key Technologies for Data-Driven 3D Shape Generation and Completion. [Ph.D. Thesis, University of Electronic Science and Technology of China]."},{"key":"ref_18","unstructured":"Hua, B. (2025). Research on Lightweight 3D SLAM Localization Technology for Mobile Robots. [Master\u2019s Thesis, University of Electronic Science and Technology of China]."},{"key":"ref_19","unstructured":"Tan, G. (2025). Research on Roadside Point Cloud Object Detection Algorithm Based on Lightweight Network. [Master\u2019s Thesis, Chongqing University of Technology]."},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Chen, Q. (2019). F-Cooper: Feature based Cooperative Perception for Autonomous Vehicle Edge Computing System Using 3D Point Clouds[EB\/OL]. arXiv.","DOI":"10.1109\/ICDCS.2019.00058"},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"342","DOI":"10.30897\/ijegeo.667393","article-title":"Calculation of Volume Changes with 3D Point Clouds of Different Physical Environments","volume":"6","author":"Acar","year":"2019","journal-title":"Int. J. Environ. Geoinform."},{"key":"ref_22","unstructured":"Ioannides, G., Chadha, A., and Elkins, A. (2024). Density adaptive attention is all you need: Robust parameter-efficient fine-tuning across multiple modalities[EB\/OL]. arXiv."},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Ruchay, A., and Fedorova, M. (2021). Fast algorithm of 3D object volume calculation from point cloud. Applications of Digital Image Processing XLIV, SPIE.","DOI":"10.1117\/12.2593891"},{"key":"ref_24","unstructured":"Lienard, J.F. (2019). Fitting 3D Shapes from Partial and Noisy Point Clouds with Evolutionary Computing[EB\/OL]. arXiv."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"1639","DOI":"10.4028\/www.scientific.net\/AMR.926-930.1639","article-title":"3D Modeling for Hillside Landscape Engineering","volume":"926\u2013930","author":"Wei","year":"2014","journal-title":"Adv. Mater. Res."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"532","DOI":"10.1109\/TLT.2016.2638811","article-title":"Point Cloud-Based Automatic Assessment of 3D Computer Animation Courseworks","volume":"10","author":"Paravati","year":"2017","journal-title":"IEEE Trans. Learn. Technol."},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Li, W.N., Piao, M.L., Jeon, S.H., Jeong, J.R., and Kim, N. (2015). Novel computer-generated hologram (CGH) achieved scheme using point cloud based on integral imaging. Advances in Display Technologies V, SPIE.","DOI":"10.1117\/12.2080714"},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"Zhao, Y., Li, G., Piao, M.L., Lee, H.M., and Kim, N. (2014). Occlusion-removed computer generated cylindrical hologram using 3D point cloud. Practical Holography XXVIII: Materials and Applications, SPIE.","DOI":"10.1117\/12.2039775"},{"key":"ref_29","first-page":"1","article-title":"Research on Bulk Grain Pile Volume Measurement Method Based on DUSt3R Stereoscopic Deep Learning","volume":"46","author":"Chen","year":"2025","journal-title":"J. Henan Univ. Technol. (Nat. Sci. Ed.)"}],"container-title":["Information"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2078-2489\/17\/2\/184\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,11]],"date-time":"2026-02-11T11:34:39Z","timestamp":1770809679000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2078-2489\/17\/2\/184"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,11]]},"references-count":29,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2026,2]]}},"alternative-id":["info17020184"],"URL":"https:\/\/doi.org\/10.3390\/info17020184","relation":{},"ISSN":["2078-2489"],"issn-type":[{"value":"2078-2489","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,11]]}}}