Computers & Graphics-Uk最新文献

{"title":"Menu user interface design for Head-mounted display Virtual Reality: A systematic literature review","authors":"Rishabh Pandey,&nbsp;Keyur Sorathia","doi":"10.1016/j.cag.2025.104402","DOIUrl":"10.1016/j.cag.2025.104402","url":null,"abstract":"<div><div>Head-mounted display (HMD) Virtual Reality (VR) has seen significant developments in recent years. The tech is finally being adopted by millions of consumers worldwide. System control through menu interfaces is crucial in VR for task completion. Unlike traditional two-dimensional (2D) devices, HMD-VR allows six degrees of freedom (DoF) interaction and navigation in three-dimensional (3D) and 360°virtual environments (VEs). This makes designing and placing user interface (UI) elements in VEs challenging because there is no apparent surface or boundary in VE to place UI elements like menus. Over the recent years, VR app stores have seen the introduction of numerous VR applications. However, there is a notable lack of guidance for designers and developers on designing and integrating crucial UI components, such as menus, for HMD-VR applications. This paper presents a systematic literature review of HMD-VR menu UIs. A total of 33 studies from Scopus and Web of Science databases published in English between 2012–2024 were identified using the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. Results describe (a) key factors that are considered when designing and evaluating menus for HMD-VR, (b) state-of-the-art on types of menu UI based on integration, and (c) study-specific design guidelines proposed on menus for HMD-VR. Furthermore, we propose synthesized, context-aware recommendations to inform future menu UI development. By bridging research insights with practical design implications, this review aims to support the creation of more usable, immersive, and goal-oriented VR interfaces.</div></div>","PeriodicalId":50628,"journal":{"name":"Computers & Graphics-Uk","volume":"132 ","pages":"Article 104402"},"PeriodicalIF":2.8,"publicationDate":"2025-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145003717","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":"Unsupervised image smoothing framework with multi-scale separated convolutional attention and multiple loss constraints","authors":"Jiafu Zeng ,&nbsp;Yepeng Liu ,&nbsp;Huiyu Li ,&nbsp;Zhaolong Gao","doi":"10.1016/j.cag.2025.104413","DOIUrl":"10.1016/j.cag.2025.104413","url":null,"abstract":"<div><div>Image smoothing is a fundamental task in image processing, aiming to remove fine textures as much as possible while preserving the primary structural information. Currently, the effectiveness of traditional methods depends heavily on handcrafted priors and manual parameter tuning. Deep learning methods have enabled automatic optimization, but they typically require paired datasets for supervision. To overcome these challenges, this paper proposed an unsupervised image smoothing framework that combines attention mechanisms and multiple loss constraints. It adopts cycle-consistent generative adversarial network structure, suitable for unpaired data. To support training, we firstly constructed an unpaired dataset and proposed an image smoothing algorithm based on image decomposition and multi-scale total variation. Secondly, we proposed a Multi-scale Separated Convolutional Attention (MSCA) module, which adaptively captures and fuses spatial and channel-wise information across scales. Additionally, we incorporated two loss terms: a perceptual cycle-consistency loss using learned perceptual image patch similarity to measure high-level feature similarity, and a total variation loss to promote local smoothness in the generated images. The experimental results can demonstrate that the proposed method achieves good results in both qualitative and quantitative evaluations, comparable to other state-of-the-art smoothing methods. Compared to other unsupervised methods, it exhibits superior performance and broader applications.</div></div>","PeriodicalId":50628,"journal":{"name":"Computers & Graphics-Uk","volume":"132 ","pages":"Article 104413"},"PeriodicalIF":2.8,"publicationDate":"2025-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145010134","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 faster single-image denoising diffusion model: Emphasizing the role of the latent image code","authors":"Can Bai ,&nbsp;Xianjun Han","doi":"10.1016/j.cag.2025.104329","DOIUrl":"10.1016/j.cag.2025.104329","url":null,"abstract":"<div><div>Stable diffusion models have produced remarkable advancements in image generation tasks, which suggests that diffusion processes are conducted in the latent space. However, approaches of this type rely heavily on extensive text–image datasets for training purposes. Various methods train on and extract the intrinsic characteristics of a single image to generate various images that conform to the distribution of the training image. Nevertheless, training directly in the image space is time-consuming. In this work, we investigate the use of the latent space representation of a single image to generate images by iteratively employing a lightweight diffusion model. In particular, we integrate the highly realistic.</div><div>VQGAN and a variational autoencoder (VAE) model into a new encoder to obtain the latent code of the input image; this enables pixel-to-latent space transformations and fast training to form high-resolution images by learning a latent representation of the input image. After obtaining the latent code, we apply lightweight diffusion models that remove the standard upsampling and downsampling layers to reconstruct the latent code. The decoder receives the latent code generated through sampling and converts it to the pixel space, thereby generating high-quality images that follow the distribution of the training image. This design enables us to perceive a more realistic distribution for the input image and accelerate the inference and sampling processes. Compared with existing pixel-based single-image diffusion models, our method greatly accelerates the training and inference procedures while ensuring the quality and diversity of the generated samples.</div></div>","PeriodicalId":50628,"journal":{"name":"Computers & Graphics-Uk","volume":"132 ","pages":"Article 104329"},"PeriodicalIF":2.8,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144913862","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":"Generation of semantic isomorphic mesh using implicit function representation","authors":"Shoko Miyauchi ,&nbsp;Kyo Itaya ,&nbsp;Ken’ichi Morooka","doi":"10.1016/j.cag.2025.104407","DOIUrl":"10.1016/j.cag.2025.104407","url":null,"abstract":"<div><div>We propose a novel point-to-mesh network for generating isomorphic meshes (IMs) from point clouds of different classes and semantic IMs (SIMs) from point clouds within the same class. IMs represent arbitrary objects with a unified mesh structure. SIMs ensure semantic correspondence among IMs of the same class. Here, the semantic correspondence means that common feature points of different IMs in the same class are represented by the vertices with the same identification numbers while satisfying a one-to-one vertex correspondence among the meshes. The unified mesh structure of IMs enables the same process to be applied to all meshes. Moreover, using SIMs can ensure efficient memory usage and reduce calculation time in mesh-targeted applications while maintaining their accuracy. In the proposed method, for each class, a template shape, which has shape features common among the objects belonging to the class, is trained using implicit function representation. Subsequently, a template mesh for each class is obtained by describing the template shape using the mesh structure of a previously prepared reference mesh. In inference processing, the input point cloud of each class is described by the mesh structure of the reference mesh by performing an inverse mapping from the template mesh of the class to the input three-dimensional object. Through the SIM generations from point clouds with/without noise or missing parts, we confirmed that the proposed method is capable of highly accurate shape restoration with low computational cost compared to conventional methods.</div></div>","PeriodicalId":50628,"journal":{"name":"Computers & Graphics-Uk","volume":"132 ","pages":"Article 104407"},"PeriodicalIF":2.8,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144922109","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":"Beyond buttons: A user-centric approach to hands-free locomotion in Virtual Reality via voice commands","authors":"Jan Hombeck,&nbsp;Henrik Voigt,&nbsp;Kai Lawonn","doi":"10.1016/j.cag.2025.104318","DOIUrl":"10.1016/j.cag.2025.104318","url":null,"abstract":"<div><div>Using speech for hands-free interaction in Virtual Reality (VR) is gaining popularity, supported by advances in natural language processing that enable accurate speech-to-text transcription and intent inference. Prior work has shown that voice-based navigation is effective when hand use is restricted or occupied, particularly when visually salient points in the scene, known as landmarks, can be identified and articulated. However, in many virtual environments, such landmarks may be absent or not easily recognizable, limiting the applicability of voice commands. To address this, we propose three landmark-free, user-centered coordinate systems to support speech-based locomotion in VR: Cartesian, cylindrical, and spherical. Each system uses a distinct encoding and interaction style. The Cartesian system employs a simple three-digit code and requires low cognitive effort. The spherical system prioritizes precision but demands higher mental effort. The cylindrical system combines elements of both, offering a balance between usability and accuracy. We evaluated these systems in a quantitative user study with 24 participants, all with backgrounds in information technology. The study assessed the systems’ effectiveness for object positioning and navigation, comparing them to teleportation, the standard non-voice locomotion method in VR. Results show that the Cartesian system enables faster and more intuitive navigation compared to the spherical system. Distinct usage patterns were observed, with users predominantly focusing on the central field of view during navigation. Our findings indicate that user-centered coordinate systems are practical to implement and present a viable alternative for speech-driven, hands-free navigation in VR.</div></div>","PeriodicalId":50628,"journal":{"name":"Computers & Graphics-Uk","volume":"132 ","pages":"Article 104318"},"PeriodicalIF":2.8,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144989148","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":"Uniform parametric mapping of saccular intracranial aneurysms for statistical analysis of morphological variation","authors":"Pepe Eulzer,&nbsp;Kai Lawonn","doi":"10.1016/j.cag.2025.104408","DOIUrl":"10.1016/j.cag.2025.104408","url":null,"abstract":"<div><div>The morphology of saccular intracranial aneurysms is widely assumed to encode biomechanical information critical for rupture risk, yet existing predictors rely almost exclusively on handcrafted shape descriptors that are subjective and capture only limited aspects of the geometry. We introduce a data-driven framework for uniform parametric mapping and statistical shape analysis of aneurysm surfaces, applied to a multi-institutional set of 958 patient-extracted intracranial aneurysms. We develop a uniform parametric mapping based on an angular- and radial-field parameterization, yielding a surface map into a registered canonical space. The mapping is fully automatic, requires no manual annotations, and is independent of the cut configuration, meshing regularity, and sampling density. Using the mapping, we create uniform point correspondences across all aneurysms. From these correspondences, high-dimensional shape vectors are constructed, and principal component analysis (PCA) yields the mean aneurysm shape and dominant modes of variation. We investigate how strongly each mode contributes to the rupture association using logistic regression of the PCA coefficients. We identified three modes significantly associated with rupture status (<span><math><mrow><mi>p</mi><mo>&lt;</mo><mn>0</mn><mo>.</mo><mn>01</mn></mrow></math></span>) and provide interpretable deformation patterns for their morphological characteristics. The shape-based regression model classifying rupture status achieves an average accuracy of <span><math><mrow><mn>0</mn><mo>.</mo><mn>76</mn><mo>±</mo><mn>0</mn><mo>.</mo><mn>02</mn></mrow></math></span> and an AUC of <span><math><mrow><mn>0</mn><mo>.</mo><mn>77</mn><mo>±</mo><mn>0</mn><mo>.</mo><mn>01</mn></mrow></math></span> after 5-fold cross-validation. Based on a diverse range of aneurysm models, our framework demonstrates robust parameter mapping and statistical measures, offering a generalized, reproducible approach for shape-based risk stratification that may inform evidence-based management of unruptured aneurysms.</div></div>","PeriodicalId":50628,"journal":{"name":"Computers & Graphics-Uk","volume":"132 ","pages":"Article 104408"},"PeriodicalIF":2.8,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144917626","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":"SHREC 2025: Protein surface shape retrieval including electrostatic potential","authors":"Taher Yacoub ,&nbsp;Camille Depenveiller ,&nbsp;Atsushi Tatsuma ,&nbsp;Tin Barisin ,&nbsp;Eugen Rusakov ,&nbsp;Udo Göbel ,&nbsp;Yuxu Peng ,&nbsp;Shiqiang Deng ,&nbsp;Yuki Kagaya ,&nbsp;Joon Hong Park ,&nbsp;Daisuke Kihara ,&nbsp;Marco Guerra ,&nbsp;Giorgio Palmieri ,&nbsp;Andrea Ranieri ,&nbsp;Ulderico Fugacci ,&nbsp;Silvia Biasotti ,&nbsp;Ruiwen He ,&nbsp;Halim Benhabiles ,&nbsp;Adnane Cabani ,&nbsp;Karim Hammoudi ,&nbsp;Matthieu Montes","doi":"10.1016/j.cag.2025.104394","DOIUrl":"10.1016/j.cag.2025.104394","url":null,"abstract":"<div><div>This SHREC 2025 track dedicated to protein surface shape retrieval involved 9 participating teams. We evaluated the performance in retrieval of 15 proposed methods on a large dataset of 11,555 protein surfaces with calculated electrostatic potential (a key molecular surface descriptor). The performance in retrieval of the proposed methods was evaluated through different metrics (Accuracy, Balanced accuracy, F1 score, Precision and Recall). The best retrieval performance was achieved by the proposed methods that used the electrostatic potential complementary to molecular surface shape. This observation was also valid for classes with limited data which highlights the importance of taking into account additional molecular surface descriptors.</div></div>","PeriodicalId":50628,"journal":{"name":"Computers & Graphics-Uk","volume":"132 ","pages":"Article 104394"},"PeriodicalIF":2.8,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145003715","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":"Orientation control via G3 parametric quaternion interpolation spline curves","authors":"Qian Wang ,&nbsp;Jiayi Zhu ,&nbsp;Aihua Liang ,&nbsp;Mengyao Zhang ,&nbsp;Junxiao Xue","doi":"10.1016/j.cag.2025.104383","DOIUrl":"10.1016/j.cag.2025.104383","url":null,"abstract":"<div><div>In order to interpolate the real rigid body motion orientation, this paper proposes a unit quaternion spline curve based on the parametric 6th-degree polynomial in Euclidean space. The quaternion spline curve preserves <span><math><msup><mrow><mi>G</mi></mrow><mrow><mn>3</mn></mrow></msup></math></span> continuity and interpolates a given sequence of points; meanwhile, it possesses the local shape adjustability. Through numerical experiments, the influence of parameter variation on motion orientation is specifically analyzed and the performance of the method in the case of a sudden rotation is also considered.</div></div>","PeriodicalId":50628,"journal":{"name":"Computers & Graphics-Uk","volume":"132 ","pages":"Article 104383"},"PeriodicalIF":2.8,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144926384","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":"Point cloud segmentation for 3D Clothed Human Layering","authors":"Davide Garavaso ,&nbsp;Federico Masi ,&nbsp;Pietro Musoni ,&nbsp;Umberto Castellani","doi":"10.1016/j.cag.2025.104393","DOIUrl":"10.1016/j.cag.2025.104393","url":null,"abstract":"<div><div>3D Cloth modeling and simulation is essential for avatars creation in several fields, such as fashion, entertainment, and animation. Achieving high-quality results is challenging due to the large variability of clothed body especially in the generation of realistic wrinkles. 3D scan acquisitions provide more accuracy in the representation of real-world objects but lack semantic information that can be inferred with a reliable semantic reconstruction pipeline. To this aim, shape segmentation plays a crucial role in identifying the semantic shape parts. However, current 3D shape segmentation methods are designed for scene understanding and interpretation and only few work is devoted to modeling. In the context of clothed body modeling the segmentation is a preliminary step for fully semantic shape parts reconstruction namely the underlying body and the involved garments. These parts represent several layers with strong overlap in contrast with standard segmentation methods that provide disjoint sets. In this work we propose a new 3D point cloud segmentation paradigm where each 3D point can be simultaneously associated to different layers. In this fashion we can estimate the underlying body parts and the unseen clothed regions, i.e., the part of a cloth occluded by the clothed-layer above. We name this segmentation paradigm <em>clothed human layering</em>. We create a new synthetic dataset that simulates very realistic 3D scans with the ground truth of the involved clothing layers. We propose and evaluate different neural network settings to deal with 3D clothing layering. We considered both coarse and fine grained per-layer garment identification. Our experiments demonstrates the benefit in introducing proper strategies for the segmentation on the garment domain on both the synthetic and real-world scan datasets.</div></div>","PeriodicalId":50628,"journal":{"name":"Computers & Graphics-Uk","volume":"132 ","pages":"Article 104393"},"PeriodicalIF":2.8,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144913730","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":"Perspective scale breaks: A user study and proposal to mitigate the impact of outliers in bar chart plots","authors":"Natã Ferreira Lobato,&nbsp;Kevin Washington Azevedo da Cruz,&nbsp;Bianchi Serique Meiguins,&nbsp;Carlos Gustavo Resque dos Santos","doi":"10.1016/j.cag.2025.104390","DOIUrl":"10.1016/j.cag.2025.104390","url":null,"abstract":"<div><div>Bar charts are widely used for interpreting quantitative data, but they can be challenging to use when dealing with datasets that contain outliers owing to limited visual space. This study aims to assess techniques for addressing this issue and proposes an approach called the perspective scale break. The perspective scale break introduces a perspective effect, providing a focus+context view and effectively addressing datasets with outliers. We conducted a user study that combined quantitative measures (accuracy and task completion time) and qualitative insights (obtained through participant interviews) to assess the effectiveness of the proposed method. We compared the performance of the Perspective Scale Break against default bar chart variations, including default scaling, logarithmic scaling, and traditional scale breaks, across a range of dataset types and analytical tasks. Our study determined the most effective bar chart technique based on the dataset and task. Additionally, we explored participants’ preferences to identify which charts were easiest to interpret, most visually appealing, and most difficult to use.</div></div>","PeriodicalId":50628,"journal":{"name":"Computers & Graphics-Uk","volume":"132 ","pages":"Article 104390"},"PeriodicalIF":2.8,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144926383","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}