{"title":"Efficient neural RGB-D indoor scene reconstruction based on normal features","authors":"Xiaoqun Wu, Xin Liu, Yumeng Cao, Haisheng Li","doi":"10.1016/j.cagd.2025.102452","DOIUrl":"10.1016/j.cagd.2025.102452","url":null,"abstract":"<div><div>Reconstructing large-scale indoor scenes from 2D images to 3D models presents substantial challenges, particularly in handling texture-less regions and extensive scene sizes with both accuracy and efficiency. This paper introduces a novel method for efficient and high-quality geometric reconstruction of indoor scenes using RGB-D images. Our approach integrates normal features as prior information into the RGB-D data and employs a truncated signed distance function (TSDF) to represent scene surfaces. Combined with multi-resolution hash encoding, the proposed method achieves both high reconstruction quality and computational efficiency. Specifically, we estimate normal vectors from RGB images as feature priors to guide surface fitting. To address the inaccuracies of normal estimation in regions with small objects or complex geometric details, we incorporate depth information to better constrain the surface fitting process. Additionally, multi-resolution hash encoding is used to stratify sampling points, enabling rapid feature lookups via hash functions. Experimental results demonstrate that the proposed method significantly outperforms existing approaches in terms of both reconstruction quality and computational efficiency.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"119 ","pages":"Article 102452"},"PeriodicalIF":1.3,"publicationDate":"2025-04-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143882631","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"What smooth surfaces can be constructed from total degree 2 splines?","authors":"Jörg Peters, Kȩstutis Karčiauskas","doi":"10.1016/j.cagd.2025.102435","DOIUrl":"10.1016/j.cagd.2025.102435","url":null,"abstract":"<div><div>On a planar Euclidean domain, Powell-Sabin splines form a rich space of <span><math><msup><mrow><mi>C</mi></mrow><mrow><mn>1</mn></mrow></msup></math></span> polynomials of total degree 2, i.e. with constant second derivatives. However, when the domain has a different structure because the genus of the surface is not 1, building curved free-form surfaces solely with total degree quadratic polynomials, with each piece defined over a flat, straight-edge domain triangle, meets with obstructions. By pinpointing these obstructions, the limitations of modeling with quadratics are made precise, the allowable <span><math><msup><mrow><mi>C</mi></mrow><mrow><mn>1</mn></mrow></msup></math></span> free-form constructions are characterized and their necessary shape-deficiency is demonstrated.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"119 ","pages":"Article 102435"},"PeriodicalIF":1.3,"publicationDate":"2025-04-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143886732","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Feature-preserving point cloud filtering via mixture family manifold","authors":"Peng Du, Xingce Wang, Yaohui Fang, Xudong Ru, Haichuan Zhao, Zhongke Wu","doi":"10.1016/j.cagd.2025.102453","DOIUrl":"10.1016/j.cagd.2025.102453","url":null,"abstract":"<div><div>Filtering noisy point cloud of complex models while effectively preserving geometric features, especially fine-scale features, presents the main challenge. In this paper, we propose a non-learning, feature-preserving point cloud filtering method from the novel perspective of mixture family manifold, which does not require normal estimation and does not depend on the distribution of the input data. Our novel perspective refers to formulate a potential function regularization term, related to Shannon entropy, within the mixture family manifold parameterized by the mixture weights. This regularization constrains the parameter estimation in the point cloud filtering model inspired by the Gaussian Mixture Model (GMM), avoiding the use of purely distance-based isotropic weights. Our method effectively removes noise while preserving geometric details. Experimental results on both synthetic and scanned data demonstrate that our approach outperforms the selected state-of-the-art methods, including those that roughly utilize normal information for point cloud filtering.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"119 ","pages":"Article 102453"},"PeriodicalIF":1.3,"publicationDate":"2025-04-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143886733","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Transfinite barycentric coordinates for arbitrary planar domains","authors":"Qingjun Chang, Kai Hormann","doi":"10.1016/j.cagd.2025.102433","DOIUrl":"10.1016/j.cagd.2025.102433","url":null,"abstract":"<div><div>Generalized barycentric coordinates provide a simple way of interpolating data given at the vertices of a polygon or polyhedron, with widespread applications in computer graphics, geometry processing, and other fields. Transfinite barycentric coordinates, also known as barycentric kernels, extend this idea to curved domains and can be used to interpolate continuous data given on the boundary of such domains. We present a novel framework for defining non-negative barycentric kernels over arbitrary bounded planar domains. This framework is inspired by the construction of a transfinite version of maximum likelihood coordinates and can be used to define a variety of barycentric kernels, including a simple pseudo-harmonic kernel and a non-negative variant of the mean value kernel. Moreover, we propose a novel barycentric kernel which yields transfinite interpolants that are similar to harmonic interpolants. We tested our new kernel for domains and boundary data described by closed uniform quadratic splines and in particular for image deformation. The results indicate that our method has several advantages over alternative approaches.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"119 ","pages":"Article 102433"},"PeriodicalIF":1.3,"publicationDate":"2025-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143868337","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Zhenyu Shu , Zhichao Zhang , Yiming Zhao , Teng Wu
{"title":"MTSegNet: Manifold Transformer for 3D shape segmentation","authors":"Zhenyu Shu , Zhichao Zhang , Yiming Zhao , Teng Wu","doi":"10.1016/j.cagd.2025.102440","DOIUrl":"10.1016/j.cagd.2025.102440","url":null,"abstract":"<div><div>The semantic segmentation of 3D meshes is a critical component of 3D shape analysis, which involves assigning semantic labels to each face of a 3D mesh. Despite its significance, current methods often struggle to capture manifold information in 3D meshes, a fundamental characteristic distinguishing them from other representation forms of 3D data, like 3D point clouds or 3D voxels, resulting in suboptimal segmentation outcomes. In this paper, we propose a novel Transformer-based approach, Manifold Transformer (MTSegNet), for 3D mesh semantic segmentation, which effectively learns manifold information. By using hierarchical Transformers, MTSegNet can capture both local and global features of 3D meshes, while reducing the computational complexity and memory consumption. To further improve the performance of our method, we design an effective input-generating algorithm that serializes input data into multiple sequences of tokens that represent the geometry and topology of 3D meshes. This algorithm preserves the structural information and spatial relations of 3D meshes, while enabling the use of standard Transformer architectures. The proposed method is evaluated on four benchmark datasets: PSB, COSEG, ShapeNetCore, and HumanBody, and it achieves state-of-the-art results on all datasets, outperforming the previous methods.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"119 ","pages":"Article 102440"},"PeriodicalIF":1.3,"publicationDate":"2025-04-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143894411","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Efficient worst-case topology optimization of self-supporting structures for additive manufacturing","authors":"Nan Zheng, Xiaoya Zhai, Falai Chen","doi":"10.1016/j.cagd.2025.102441","DOIUrl":"10.1016/j.cagd.2025.102441","url":null,"abstract":"<div><div>In worst-case topology optimization, uncertain loads can result in complex internal structures and intricate printed details that challenge manufacturability. However, the impact of these features on manufacturing performance is often overlooked, potentially compromising the printability and quality of the final product in additive manufacturing (AM). This paper introduces a novel approach for generating 3D self-supporting structures under worst-case topology optimization. The proposed framework utilizes an implicit tensor-product B-spline (ITPBS) representation, directly adopting its coefficients as design variables to minimize compliance while enforcing self-supporting constraints and minimal length scale. By reformulating AM constraints, we analytically derive a single geometric fabrication constraint that simultaneously addresses both overhang regions and the dripping effect. The solid-void boundary representation provided by ITPBS enables seamless integration of fabrication constraints into the worst-case optimization process. Worst-case compliance is evaluated by solving an eigenvalue problem, and sensitivity analysis is conducted using the adjoint variable method. Numerical experiments demonstrate that the proposed approach effectively produces self-supporting structures across various models.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"119 ","pages":"Article 102441"},"PeriodicalIF":1.3,"publicationDate":"2025-04-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143868338","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"On the triple-parameter least squares progressive iterative approximation and its convergence analysis","authors":"Nian-Ci Wu , Chengzhi Liu , Juncheng Li","doi":"10.1016/j.cagd.2025.102439","DOIUrl":"10.1016/j.cagd.2025.102439","url":null,"abstract":"<div><div>In the domain of large-scale data fitting, the least squares progressive iterative approximation (LSPIA) method, introduced by <span><span>Deng and Lin (2014)</span></span>, stands as a prominent figure in geometric iterative techniques due to its simplicity and effectiveness. Herein, we incorporate a triple-parameter modification to fine-tune the control points, hereby denoted as the TLSPIA method. Opting for a suitable parameter within the TLSPIA framework reverts it to the traditional LSPIA and its variant, such as PmLSPIA proposed by <span><span>Liu et al. (2024)</span></span>. Through singular value decomposition of the collocation matrix, we derive closed forms for the error sequences generated by the LSPIA, PmLSPIA, and TLSPIA methods, demonstrating that these sequences exhibit linear decrease and diverge in their convergence rates across different subspaces corresponding to any right singular vector of the collocation matrix. Empirical numerical examples are provided to substantiate our theoretical claims, highlighting the practical viability and efficacy of the proposed algorithm.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"119 ","pages":"Article 102439"},"PeriodicalIF":1.3,"publicationDate":"2025-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143860258","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"ESA-GS: Elongation splitting and assimilation in Gaussian splatting for accurate surface reconstruction","authors":"Yuyang Chen, Wenming Wu, Yusheng Peng, Yue Fei, Liping Zheng","doi":"10.1016/j.cagd.2025.102434","DOIUrl":"10.1016/j.cagd.2025.102434","url":null,"abstract":"<div><div>Recently, 3D Gaussian Splatting (3DGS) has significantly advanced the development of 3D reconstruction by providing efficient and high-quality rendering. 2D Gaussian Splatting (2DGS) introduced two-dimensional surfels as scene primitives to address 3DGS's limitations in surface representation. However, its adaptive control strategy may still result in suboptimal results, especially when dealing with extreme-shaped or large Gaussians on the surface. We propose Elongation Splitting and Assimilation in Gaussian Splatting (ESA-GS) to enhance geometric reconstruction quality by addressing these special Gaussians. Specifically, ESA-GS splits highly elongated Gaussians on the surface into three assimilated Gaussians during the densification process. In addition, ESA-GS adds an opacity degeneration strategy and an additional pruning strategy to remove invalid Gaussians and improve the geometry quality. Experimental results demonstrate that ESA-GS can produce geometrically accurate reconstructed surfaces without sacrificing efficiency in most cases.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"119 ","pages":"Article 102434"},"PeriodicalIF":1.3,"publicationDate":"2025-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143864163","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Topology guaranteed and error controlled curve tracing for parametric surface-surface intersection","authors":"Bingwei Zhang , Jin-San Cheng , Yu-Shen Liu , Zhaoqi Zhang","doi":"10.1016/j.cagd.2025.102432","DOIUrl":"10.1016/j.cagd.2025.102432","url":null,"abstract":"<div><div>We present a novel and efficient curve tracing method of the intersection of two parametric surfaces with error bound controlled and correct topology. Our method decomposes the 4D intersection curve in the parameter space into strongly monotonic curve segments such that their corresponding 3D curve segments in the model space are also strongly monotonic. This decomposition strategy can prevent straying or looping and maintain the 3D topology between the curve segments. Furthermore, by controlling the density of decomposition in the model space, our method can easily control the error bound of the numerical approximation of the curve segments. As a result, our method is very efficient compared to previous related work. Our experiments support our claims.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"118 ","pages":"Article 102432"},"PeriodicalIF":1.3,"publicationDate":"2025-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143844767","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A shift-invariant C12-subdivision algorithm which rotates the lattice","authors":"Cédric Gérot , Malcolm A. Sabin","doi":"10.1016/j.cagd.2025.102430","DOIUrl":"10.1016/j.cagd.2025.102430","url":null,"abstract":"<div><div>In order to study differential properties of a subdivision surface at a markpoint, it is necessary to parametrise it over a so-called characteristic map defined as the infinite union of <span><math><msup><mrow><mi>C</mi></mrow><mrow><mi>k</mi></mrow></msup></math></span>-parametrised rings. Construction of this map is known when a single step of the subdivision scheme does not rotate a regular lattice. Otherwise, two steps are considered as they realign the lattice and its subdivided version. We present a new subdivision scheme which rotates the lattice and nevertheless allows a direct construction of the characteristic map. It is eigenanalysed with techniques introduced in a companion article and proved to define a <span><math><msubsup><mrow><mi>C</mi></mrow><mrow><mn>1</mn></mrow><mrow><mn>2</mn></mrow></msubsup></math></span>-algorithm around a face-centre. This scheme generalises Loop's scheme, allowing the designer to choose between extraordinary vertices or faces in regard to the shape of the mesh, the location of the extraordinary elements, and the aimed limit shape.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"118 ","pages":"Article 102430"},"PeriodicalIF":1.3,"publicationDate":"2025-03-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143714698","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}