Shape Classification Research Articles

RJAN: Region‐based joint attention network for 3D shape recognition

AbstractAs an essential field of multimedia and computer vision, 3D shape recognition has attracted much research attention in recent years. Multiview‐based approaches have demonstrated their superiority in generating effective 3D shape representations. Typical methods usually extract the multiview global features and aggregate them together to generate 3D shape descriptors. However, there exist two disadvantages: First, the mainstream methods ignore the comprehensive exploration of local information in each view. Second, many approaches roughly aggregate multiview features by adding or concatenating them together. The information loss for some discriminative characteristics limits the representation effectiveness. To address these problems, a novel architecture named region‐based joint attention network (RJAN) was proposed. Specifically, the authors first design a hierarchical local information exploration module for view descriptor extraction. The region‐to‐region and channel‐to‐channel relationships from different granularities can be comprehensively explored and utilised to provide more discriminative characteristics for view feature learning. Subsequently, a novel relation‐aware view aggregation module is designed to aggregate the multiview features for shape descriptor generation, considering the view‐to‐view relationships. Extensive experiments were conducted on three public databases: ModelNet40, ModelNet10, and ShapeNetCore55. RJAN achieves state‐of‐the‐art performance in the tasks of 3D shape classification and 3D shape retrieval, which demonstrates the effectiveness of RJAN. The code has been released on https://github.com/slurrpp/RJAN.

FSH3D: 3D Representation via Fibonacci Spherical Harmonics

AbstractSpherical harmonics are a favorable technique for 3D representation, employing a frequency‐based approach through the spherical harmonic transform (SHT). Typically, SHT is performed using equiangular sampling grids. However, these grids are non‐uniform on spherical surfaces and exhibit local anisotropy, a common limitation in existing spherical harmonic decomposition methods. This paper proposes a 3D representation method using Fibonacci Spherical Harmonics (FSH3D). We introduce a spherical Fibonacci grid (SFG), which is more uniform than equiangular grids for SHT in the frequency domain. Our method employs analytical weights for SHT on SFG, effectively assigning sampling errors to spherical harmonic degrees higher than the recovered band‐limited function. This provides a novel solution for spherical harmonic transformation on non‐equiangular grids. The key advantages of our FSH3D method include: 1) With the same number of sampling points, SFG captures more features without bias compared to equiangular grids; 2) The root mean square error of 32‐degree spherical harmonic coefficients is reduced by approximately 34.6% for SFG compared to equiangular grids; and 3) FSH3D offers more stable frequency domain representations, especially for rotating functions. FSH3D enhances the stability of frequency domain representations under rotational transformations. Its application in 3D shape reconstruction and 3D shape classification results in more accurate and robust representations. Our code is publicly available at https://github.com/Miraclelzk/Fibonacci-Spherical-Harmonics.

Neural Vector Fields for Implicit Surface Representation and Inference

AbstractNeural implicit fields have recently shown increasing success in representing, learning and analysis of 3D shapes. Signed distance fields and occupancy fields are still the preferred choice of implicit representations with well-studied properties, despite their restriction to closed surfaces. With neural networks, unsigned distance fields as well as several other variations and training principles have been proposed with the goal to represent all classes of shapes. In this paper, we develop a novel and yet a fundamental representation considering unit vectors in 3D space and call it Vector Field (VF). At each point in $$\mathbb {R}^3$$ R 3 , VF is directed to the closest point on the surface. We theoretically demonstrate that VF can be easily transformed to surface density by computing the flux density. Unlike other standard representations, VF directly encodes an important physical property of the surface, its normal. We further show the advantages of VF representation, in learning open, closed, or multi-layered surfaces. We show that, thanks to the continuity property of the neural optimization with VF, a separate distance field becomes unnecessary for extracting surfaces from the implicit field via Marching Cubes. We compare our method on several datasets including ShapeNet where the proposed new neural implicit field shows superior accuracy in representing any type of shape, outperforming other standard methods. Codes are available at https://github.com/edomel/ImplicitVF.

Classification of Deformable Smooth Shapes Through Geodesic Flows of Diffeomorphisms

Classification of Deformable Smooth Shapes Through Geodesic Flows of Diffeomorphisms

Optimisation of Active Magnetic Elements in Beam-like Structures—Numerical Modelling Studies

This paper explores integrating advanced materials, including magnetic shape memory alloys, magnetorheological fluids, and classical shape memory alloys, within structural elements to achieve exceptional physical properties. When these materials are integrated within structures—whether as wires, actuators, or dampers—they provide the structures with unique static, dynamic, and damping characteristics not commonly found in nature. This study aimed to evaluate the efficacy of these active materials in enhancing the performance of beam-like structures. This investigation was conducted through a comprehensive numerical analysis, focusing on a composite beam. The study examined the impact of different active elements, their position within the structure, and their influence on key dynamic properties. Additionally, a simplified damage scenario was considered, wherein the adverse effects of structural damage were mitigated through the strategic application of these materials. Numerical simulations were carried out using the finite element method, with custom computational codes developed in MATLAB. The findings of these simulations are presented and discussed in this paper.

Soil water repellency under Eucalyptus stands of different age and its relationship with soil hydrological function and soil organic matter

Eucalyptus stands can modify soil hydrological functioning, by several mechanisms including the effect of the growing root system, changes in soil organic matter (SOM) and the occurrence of soil water repellency (SWR). This effect is expected to evolve with time, as tree root system develop and SOM is formed from tree litter. SWR can prevent water from entering the soil, reducing infiltration, while the effect of roots and SOM can improve structure, favoring infiltration. The aims of this study were: i- to determine the influence of Eucalyptus stands of different ages on soil physical and hydraulic properties and SWR phenomena; ii- to evaluate the effects of SOM content and composition on SWR; iii- to determine the relationship between SWR and soil physical properties; iv- to compare different methodologies for SWR evaluation. The soil studied was a in a silty loam Typic Argiudoll with Eucalyptus stands of 3, 7, 11 and 32 years. Samples were taken to determine saturated and near saturation hydraulic conductivity, water retention curve, total organic carbon and different fractions of SOM, and SWR. Different parameters were used to estimate SWR, including the repellency index (RI), water repellency cessation time (WRCT), modified repellency index (RIm), and a qualitative analysis of the shapes of cumulative infiltration curves. As main results, a reduction in total porosity (TP) in the first years of Eucalyptus plantation was found, followed by a steep increased in TP, mainly due to an increase in microporosity. Hydraulic conductivity decreased with stand age, which was attributed to a decrease in macro and mesoporosity and porosity connectivity. SOM and most SOM fractions were not affected by stand age. RI was a consistent method to estimate SWR in Eucalyptus stands, while WRCT and RIm determination was more difficult when non classical shapes of infiltration curves were obtained.

Prevalence, impact and associated factors of abnormal preoperative investigation result in patients undergoing surgery in Dilla University referral hospital: cross-sectional study

Background: As one domain of preoperative assessment, preoperative investigations are often ordered to evaluate patient’s medical condition for risk stratification and assessing patient status to undergoing surgery. Despite the fact that laboratory testing can assist in ensuring the best possible preoperative condition, routine screening examinations have a number of drawbacks. Although there are evidence-based recommendations for which investigations should be done, the tradition of routine preoperative testing is still prevalent and clinical practice with abnormal results detected varies. Method: Institution-based cross-sectional study design was conducted from 1 November to January at Dilla University Referral Hospital. Data was collected from complete pre-anaesthesia check-up sheets, investigations already done. It was collected at the individual level by using, closed-ended self-guided questionnaire. The collected data was entered, cleaned, edited and checked using SPSS version 26 for data processing and analysis. Logistic regression was performed to examine the impacts of abnormal preoperative investigation results and summarised by using tables and figures. An Adjusted odds ratio with 95% CI was computed to determine the level of significance. Result: Data of 208 patients (65.9 female) with mean±standard deviation age 30.83±15.340 years and 22.59±2.99 BMI were analysed. Patients were mostly American Society of Anaesthesiologists I and II underwent National Institute of Clinical and Health Excellence Grade 2 surgeries and surgical shape class 3. Totally, 178 (44.5%) test results were abnormal. CBC is the most detected abnormal result. Only 15 (3.75%) abnormalities had an impact in terms of delay, further investigations, and surgical technique. Comorbidity (AOR 7.982, 95% CI, P=0.041), medication history (AOR 1.463, 95% CI, P=0.013), ASA physical status II (AOR 3.287, 95% CI, P=0.029) and history of smoking (AOR 1.577, 95% CI, P=0.049) were factors which was significantly associated with abnormal preoperative investigation result. Conclusion: Only 0.6% of all tests had a significant impact in terms of changing perioperative anaesthetic management. The significant impact of abnormal investigation result noticed was delayed surgery

Enhancing MeshNet for 3D shape classification with focal and regularization losses

With the development of deep learning and computer vision, an increasing amount of research has focused on applying deep learning models to the recognition and classification of three-dimensional shapes. In classification tasks, differences in sample quantity, feature amount, model complexity, and other aspects among different categories of 3D model data cause significant variations in classification difficulty. However, simple cross-entropy loss is generally used as the loss function, but it is insufficient to address these differences. In this paper, we used MeshNet as the base model and introduced focal loss as a metric for the loss function. Additionally, to prevent deep learning models from developing a preference for specific categories, we incorporated regularization loss. The combined use of focal loss and regularization loss in optimizing the MeshNet model’s loss function resulted in a classification accuracy of up to 92.46%, representing a 0.20% improvement over the original model’s highest accuracy of 92.26%. Furthermore, the average accuracy over the final 50 epochs remained stable at a higher level of 92.01%, reflecting a 0.71% improvement compared to the original MeshNet model’s 91.30%. These results indicate that our method performs better in 3D shape classification task.

Classification of land lot shapes in real estate sector using a convolutional neural network

In the agriculture and real estate industries, land lot shapes have mostly been classified by visual inspection or hard-crafted rules. These conventional methods are time-consuming, resource-intensive, and subject to human bias. This study aims to fill this gap and alleviate problems inherent in traditional lot classification approaches. This study attempts to classify lot shapes automatically, using a convolutional neural network. A study area was chosen, image data of the lots in the area were collected and preprocessed, and an Xception neural network was specified to classify land lots according to their shapes. The test applied to a different area adjacent to the study area achieved an accuracy of 90.1% and area under the curve (AUC) of 0.96. Additionally, this study demonstrated that shape regularity can be quantified using the output scores from the neural network analysis. This is the first attempt to employ a deep learning algorithm for land management on a micro-spatial scale. The classification approach proposed in this study is expected to encourage the rapid and accurate classification of various lot shapes.

Anisakis extracellular vesicles elicit immunomodulatory and potentially tumorigenic outcomes on human intestinal organoids

BackgroundAnisakis spp. are zoonotic nematodes causing mild to severe acute and chronic gastrointestinal infections. Chronic anisakiasis can lead to erosive mucosal ulcers, granulomas and inflammation, potential tumorigenic triggers. How Anisakis exerts its pathogenic potential through extracellular vesicles (EVs) and whether third-stage infective larvae may favor a tumorigenic microenvironment remain unclear.MethodsHere, we investigated the parasite's tumorigenic and immunomodulatory capabilities using comparative transcriptomics, qRT-PCR and protein analysis with multiplex ELISA on human intestinal organoids exposed to Anisakis EVs. Moreover, EVs were characterized in terms of shape, size and concentration using classic TEM, SEM and NTA analyses and advanced interferometric NTA.ResultsAnisakis EVs showed classic shape features and a median average diameter of around 100 nm, according to NTA and iNTA. Moreover, a refractive index of 5–20% of non-water content suggested their effective biological cargo. After treatment of human intestinal organoids with Anisakis EVs, an overall parasitic strategy based on mitigation of the immune and inflammatory response was observed. Anisakis EVs impacted gene expression of main cytokines, cell cycle regulation and protein products. Seven key genes related to cell cycle regulation and apoptosis were differentially expressed in organoids exposed to EVs. In particular, the downregulation of EPHB2 and LEFTY1 and upregulation of NUPR1 genes known to be associated with colorectal cancer were observed, suggesting their involvement in tumorigenic microenvironment. A statistically significant reduction in specific mediators of inflammation and cell-cycle regulation from the polarized epithelium as IL-33R, CD40 and CEACAM1 from the apical chambers and IL-1B, GM-CSF, IL-15 and IL-23 from both chambers were observed.ConclusionsThe results here obtained unravel intestinal epithelium response to Anisakis EVs, impacting host’s anthelminthic strategies and revealing for the first time to our knowledge the host-parasite interactions in the niche environment of an emerging accidental zoonosis. Use of an innovative EV characterization approach may also be useful for study of other helminth EVs, since the knowledge in this field is very limited.Graphical

3D characteristic indicator analysis and comprehensive evaluation of manufactured sand particles used in high-performance concrete

Two-dimensional particle morphology indices face challenges in accurately depicting the characteristics of three-dimensional (3D) manufactured sand (MS), which does not effectively guide the selection of MS for high-performance concrete (HPC). This study aimed to obtain a quantified comprehensive evaluation of actual MS particle groups with specific gradations by thoroughly analysing 3D geometric characteristics. Two representative MS samples were selected for exploration, and river sand (RS) was used as the control. A random selection of sand particles was screened on six sieve levels ranging from 0.15 mm to 4.75 mm. Using X-ray computed tomography and associated image processing technologies, 3D models of these particles were generated, and MATLAB was used to extract their 3D morphology information. Subsequently, the change rules and validity of current 3D particle morphology indicators were examined, focusing on shape, angularity, and texture. Moreover, a comprehensive morphology index, Cm, was proposed by considering the effects of gradation and particle shape classification. The results showed that shape indices more accurately reflected particle characteristics, notably, a higher proportion of spheroid particles, a lower aspect ratio, and greater sphericity correlated with improved particle shape; contrary to traditional beliefs, the angularity index suggested that greater roundness did not necessarily equate to more rounded particles owing to the influence of the proportion of spheroid and cubic particles; meanwhile, texture indices showed minimal difference in the mean values of the 3D fractal dimension for the three types of sand. The proposed index Cm effectively differentiated the particle morphology states of the three sands, with higher Cm values indicating superior particle morphology. The smaller the difference between the Cm values of MS and RS, the better substitute of MS for RS.

On Robust Stabilization of Motion of a Quadrotor with Slung Load

One of the important tasks that can be performed by quadrotors is the transportation of various payloads. If the load is suspended from the quadrotor and can move relative to the quadrotors body, this motion must be taken into account in the control system, in particular, in order to suppress and prevent unwanted oscillations of the load. In the present paper, a mechanical system is considered that consists of a quadrotor and a load suspended to it with a weightless rod. It is assumed that the cross-sectional area of the load is large enough so that the aerodynamic forces acting on it cannot be neglected. It is known that for a number of shapes of payload casing, this aerodynamic force is not reduced only to the drag force, but also contains a component orthogonal to the drag force, that is, the lift force. In order to these forces, the quasi-steady approach is used. At the same time, accurate information about the aerodynamic capabilities for each particular payload is, generally speaking, not available. However, it is known that the values of aerodynamic coefficients for a fairly wide class of body shapes are lie in a specific range. Here we discuss the problem of developing the quadrotor control that would ensure robust stabilization of the ascent and descent of the system in the conditions of incomplete information about the aerodynamic load. A method is proposed for constructing the quadrotor control for robust stabilization of uniform vertical ascent and descent of the system as a whole. It is shown that this control ensures stabilization of the target mode within a fairly wide range of system parameters. Restrictions on the target speed of quadrotor motion are determined, violation of which makes the robust stabilization impossible.

Soft cells and the geometry of seashells.

A central problem of geometry is the tiling of space with simple structures. The classical solutions, such as triangles, squares, and hexagons in the plane and cubes and other polyhedra in three-dimensional space are built with sharp corners and flat faces. However, many tilings in Nature are characterized by shapes with curved edges, nonflat faces, and few, if any, sharp corners. An important question is then to relate prototypical sharp tilings to softer natural shapes. Here, we solve this problem by introducing a new class of shapes, the soft cells, minimizing the number of sharp corners and filling space as soft tilings. We prove that an infinite class of polyhedral tilings can be smoothly deformed into soft tilings and we construct the soft versions of all Dirichlet-Voronoi cells associated with point lattices in two and three dimensions. Remarkably, these ideal soft shapes, born out of geometry, are found abundantly in nature, from cells to shells.

The Determination of Criticality for Ice Shapes Based on CCAR-25

Determining the criticality of ice shapes is a necessary condition for verifying compliance with icing airworthiness regulations. However, the clear, concise, and applicable criterion based on the geometric characteristics of ice shapes has not been clearly given out by current advisory circulars. To address this problem, this paper summarizes aerodynamic performance items and recommended ice shapes the latest version of CCAR-25 and corresponding advisory circulars for a variety of flight phases, including takeoff, holding, en route, DTO, etc., instead of the single phase of holding in the previous research. Based on the geometric classification of the ice shapes, the dominant parameters of various ice shapes are clarified by the correlation between the geometric parameters and aerodynamic effects. The geometric parameters to determine the criticality of specific ice shapes are defined as the roughness height and range for the roughness ice and the total projection height in the direction of lift for the horn ice. On this basis, the detailed determination criterion of critical ice shape geometries corresponding to different flight phases and aircraft components is formulated, which will provide an operational selection methodology for determining the geometries of critical ice shapes at the airworthiness certification stage.

Machine Learning-Based Tomato Fruit Shape Classification System.

Fruit shape significantly impacts the quality and commercial value of tomatoes (Solanum lycopersicum L.). Precise grading is essential to elucidate the genetic basis of fruit shape in breeding programs, cultivar descriptions, and variety registration. Despite this, fruit shape classification is still primarily based on subjective visual inspection, leading to time-consuming and labor-intensive processes prone to human error. This study presents a novel approach incorporating machine learning techniques to establish a robust fruit shape classification system. We trained and evaluated seven supervised machine learning algorithms by leveraging a public dataset derived from the Tomato Analyzer tool and considering the current four classification systems as label variables. Subsequently, based on class-specific metrics, we derived a novel classification framework comprising seven discernible shape classes. The results demonstrate the superiority of the Support Vector Machine model in terms of its accuracy, surpassing human classifiers across all classification systems. The new classification system achieved the highest accuracy, averaging 88%, and maintained a similar performance when validated with an independent dataset. Positioned as a common standard, this system contributes to standardizing tomato fruit shape classification, enhancing accuracy, and promoting consensus among researchers. Its implementation will serve as a valuable tool for overcoming bias in visual classification, thereby fostering a deeper understanding of consumer preferences and facilitating genetic studies on fruit shape morphometry.

Graph Transformer for 3D point clouds classification and semantic segmentation

Recently, graph-based and Transformer-based deep learning have demonstrated excellent performances on various point cloud tasks. Most of the existing graph-based methods rely on static graph, which take a fixed input to establish graph relations. Moreover, many graph-based methods apply maximizing and averaging to aggregate neighboring features, so that only a single neighboring point affects the feature of centroid or different neighboring points own the same influence on the centroid’s feature, which ignoring the correlation and difference between points. Most Transformer-based approaches extract point cloud features based on global attention and lack the feature learning on local neighbors. To solve the above issues of graph-based and Transformer-based models, we propose a new feature extraction block named Graph Transformer and construct a 3D point cloud learning network called GTNet to learn features of point clouds on local and global patterns. Graph Transformer integrates the advantages of graph-based and Transformer-based methods, and consists of Local Transformer that use intra-domain cross-attention and Global Transformer that use global self-attention. Finally, we use GTNet for shape classification, part segmentation and semantic segmentation tasks in this paper. The experimental results show that our model achieves good learning and prediction ability on most tasks. The source code and pre-trained model of GTNet will be released on https://github.com/NWUzhouwei/GTNet.

HQ-Net: A heatmap-based query backbone for point cloud understanding

With the advancement of sensor technology, including LiDAR and RGB-D cameras, tasks related to point cloud understanding have gained widespread application. The sampling and learning of feature points play a crucial role in point cloud understanding, representing a fundamental and essential aspect of 3D computer vision. In the realm of point cloud processing, traditional methods such as Farthest Point Sampling (FPS) are frequently employed to select feature points. These points are then used in conjunction with effective feature aggregation operators to facilitate reasoning. However, these traditional sampling methods primarily focus on the Euclidean distance between points, lacking a comprehensive consideration of point cloud features. To address this limitation, we introduce HQ-Net, a novel point cloud understanding model designed to optimize feature point sampling using a Heatmap-based Query Backbone. The Heatmap-based Query Backbone utilizes the Heatmap Module to select category-specific point features, effectively generating a heatmap tailored to the category features and automatically extracting features from the heatmap. Subsequently, the proposed Query Encoder is employed to facilitate the interaction between point features and query features, providing mutual guidance and enhancing the model’s ability to capture spatial structure and semantic information within point clouds. Additionally, we propose task-specific heads, such as classification and segmentation heads for different tasks, which can be seamlessly integrated with the backbone to achieve excellent performance. In various point cloud understanding tasks, including shape classification, part segmentation, and semantic segmentation, HQ-Net demonstrates excellent performance.

Individualized generation of women's prototype based on the classification of body shape

The traditional prototype method only relies on the basic dimensions of the human body such as bust girth and back length to make the prototype pattern, ignoring the differences of human body details dimensions. In order to improve the fit of garments, the research on intelligent generation technology of individualized garment patterns has become one of the hot spots in the field of clothing. Based on the 3D body scanning technology, we proposed a method for generating prototype pattern based on body shape classification. The proposed method takes as inputs the body parameters (width, thickness and angle), while the output of the method is one of the four pattern generation rules-"Y" shape, "V" shape, "I" shape, or "X" shape. 207 subjects were divided into four categories based on the body angle parameters, namely "Y" shape, "V" shape, "I" shape and "X" shape, and a shape recognition model was built. The pattern database is required to obtain pattern generation rules by 3D point cloud reconstruction and flattening. Combined with the human body shape parameters, the mathematical formulas of the feature points on the pattern landmarks are built. The accuracy of body shape recognition model is 97.4%. The error analysis of the predicted pattern parameters shows that 90% of the pattern feature points have a goodness of fit above 0.7. In the 5 main landmarks, the proportion of pattern within the absolute error range is more than 80%, indicating that the prediction effect of the pattern is good. This method can be applied to the process of automatic pattern generation system based 2D measurement to improve work efficiency.

МЕКТЕП ГЕОМЕТРИЯСЫНДАГЫ ФИГУРАЛАРДЫН КЛАССИФИКАЦИЯСЫ

This work proposes a classification of geometric figures studied in a systematic course in school geometry. Classification of geometric shapes is said to be an important aspect of teaching geometry to enable students to understand and analyze different shapes; identify similarities and differences between figures; recognize common features and differences between figures belonging to the same class or type; systematize and organize knowledge; reveal patterns and relationships between figures with similar properties, which provides the basis for subsequent discoveries; derive new properties and theorems; solve spatial geometric problems. New areas of spatial perception (virtual space, multidimensional space, digital technologies, including cloud, etc.), technologies and industries are increasingly emerging. In such changes, the relevance of the science of geometry does not disappear, but rather increases. This is due to the fact that geometric knowledge and geometric thinking give a person the basis for solving various problems in theoretical (geometric) and practical (physical) spaces. However, despite this, geometry is considered a difficult subject to perceive and understand.

Supervertex Sampling Network: A Geodesic Differential SLIC Approach for 3D Mesh.

The analysis of 3D meshes with deep learning has become prevalent in computer graphics. As an essential structure, hierarchical representation is critical for mesh pooling in multiscale analysis. Existing clustering-based mesh hierarchy construction methods involve nonlinear discretization optimization operations, making them nondifferential and challenging to embed in other trainable networks for learning. Inspired by deep superpixel learning methods in image processing, we extend them from 2D images to 3D meshes by proposing a novel differentiable chart-based segmentation method named geodesic differential supervertex (GDSV). The key to the GDSV method is to ensure that the geodesic position updates are differentiable while satisfying the constraint that the renewed supervertices lie on the manifold surface. To this end, in addition to using the differential SLIC clustering algorithm to update the nonpositional features of the supervertices, a reparameterization trick, the Gumbel-Softmax trick, is employed to renew the geodesic positions of the supervertices. Therefore, the geodesic position update problem is converted into a linear matrix multiplication issue. The GDSV method can be an independent module for chart-based segmentation tasks. Meanwhile, it can be combined with the front-end feature learning network and the back-end task-specific network as a plug-in-plug-out module for training; and be applied to tasks such as shape classification, part segmentation, and 3D scene understanding. Experimental results show the excellent performance of our proposed algorithm on a range of datasets.