Skeletonization Method Research Articles

Purity Skeleton Dynamic Hypergraph Neural Network

Recently, in the field of Hypergraph Neural Networks (HGNNs), the effectiveness of dynamic hypergraph construction has been validated, which aims to reduce structural noise within the hypergraph through embeddings. However, the existing dynamic construction methods fail to notice the reduction of information contained in the hypergraphs during dynamic updates. This limitation undermines the quality of hypergraphs. Moreover, dynamic hypergraphs are constructed from graphs. Several key nodes play a crucial role in graph, but they are overlooked in hypergraphs. In this paper, we propose a Purity Skeleton Dynamic Hypergraph Neural Network (PS-DHGNN) to address the above issues. Firstly, we leverage purity skeleton method to dynamically construct hypergraphs via the fusion embeddings of features and topology simultaneously. This method effectively reduces structural noise and prevents the loss of information. Secondly, we employ an incremental training strategy, which implements a batch training strategy based on the importance of nodes. The key nodes, as the skeleton of hypergraph, are still highly valued. In addition, we utilize a novel loss function for learning structure information between hypergraph and graph. We conduct extensive experiments on node classification and clustering tasks, which demonstrate that our PS-DHGNN outperforms state-of-the-art methods. Note on real-world traffic flow datasets, PS-DHGNN demonstrates excellent performance, which is highly meaningful in practice.

A new deep learning-based approach for concrete crack identification and damage assessment

Concrete building structures are prone to cracking as they are subjected to environmental temperatures, freeze-thaw cycles, and other operational environmental factors. Failure to detect cracks in the key building structure at the early stage can result in serious accidents and associated economic losses. A new method using the SE-U-Net model based on a conditional generative adversarial network (CGAN) has been developed to identify small cracks in concrete structures in this paper. This proposed method was a pixel-level U-Net model based on a generative network, that was integrated the original convolutional layer with an attention mechanism, and an SE module in the jump connection section was added to improve the identifiability of the model. The discriminative network compared the generated images with real images using the PatchGAN model. Through the adversarial training of generator and discriminator, the performance of generator in crack image segmentation task is improved, and the trained generation network is used to segment cracks. In damage assessments, the crack skeleton was represented by the individual pixel width and recognized using the binary morphological crack skeleton method, in which the final length, area, and average width of the crack could be determined through the geometric correction index. The results showed that compared with other methods, the proposed method could better identify subtle pixel-level cracks, and the identification accuracy is 98.48%. These methods are of great significance for the identification of cracks and the damage assessment of concrete structures in practice.

Spatial circulatory networks based on the peripheral medial axis transform

This article presents a method for constructing circulatory networks that intrinsically combine geometric, topological, and semantic spatial attributes. The objective of this work is to generate concise graph representations extracted from spatial geometry that may be used for wayfinding and code compliance analyses to inform architectural design and urban planning. Our approach is based on shape analysis using the Medial Axis Transform skeletonization method. We present a process for the semantic classification of its nodes and edges by constructing an analytical rational piece-wise 3D surface representation. This overcomes the problem of identifying salient or otherwise graph features. Furthermore, we augment the networks with peripheral pathways. This addresses the fundamental limitation of skeletal graphs forcing paths exclusively through the middle of space and thus producing unreasonable detours. We demonstrate that our approach encompasses subtle circulatory features and generates concise graphs for open and free-form spaces that do not exhibit corridor-like structures.

Quantifying robustness: 3D tree point cloud skeletonization with smart-tree in noisy domains

Extracting tree skeletons from 3D tree point clouds is challenged by noise and incomplete data. While our prior work (Dobbs et al., in: Iberian conference on pattern recognition and image analysis, Springer, Berlin, pp. 351–362, 2023) introduced a deep learning approach for approximating tree branch medial axes, its robustness against various types of noise has not been thoroughly evaluated. This paper addresses this gap. Specifically, we simulate real-world noise challenges by introducing 3D Perlin noise (to represent subtractive noise) and Gaussian noise (to mimic additive noise). To facilitate this evaluation, we introduce a new synthetic tree point cloud dataset, available at https://github.com/uc-vision/synthetic-trees-II. Our results indicate that our deep learning-based skeletonization method is tolerant to both additive and subtractive noise.

Unmanned Aerial Vehicle-Scale Weed Segmentation Method Based on Image Analysis Technology for Enhanced Accuracy of Maize Seedling Counting

The number of maize seedlings is a key determinant of maize yield. Thus, timely, accurate estimation of seedlings helps optimize and adjust field management measures. Differentiating “multiple seedlings in a single hole” of maize accurately using deep learning and object detection methods presents challenges that hinder effectiveness. Multivariate regression techniques prove more suitable in such cases, yet the presence of weeds considerably affects regression estimation accuracy. Therefore, this paper proposes a maize and weed identification method that combines shape features with threshold skeleton clustering to mitigate the impact of weeds on maize counting. The threshold skeleton method (TS) ensured that the accuracy and precision values of eliminating weeds exceeded 97% and that the missed inspection rate and misunderstanding rate did not exceed 6%, which is a significant improvement compared with traditional methods. Multi-image characteristics of the maize coverage, maize seedling edge pixel percentage, maize skeleton characteristic pixel percentage, and connecting domain features gradually returned to maize seedlings. After applying the TS method to remove weeds, the estimated R2 is 0.83, RMSE is 1.43, MAE is 1.05, and the overall counting accuracy is 99.2%. The weed segmentation method proposed in this paper can adapt to various seedling conditions. Under different emergence conditions, the estimated R2 of seedling count reaches a maximum of 0.88, with an RMSE below 1.29. The proposed approach in this study shows improved weed recognition accuracy on drone images compared to conventional image processing methods. It exhibits strong adaptability and stability, enhancing maize counting accuracy even in the presence of weeds.

A Novel Method of Skeletonization of Complex Shapes Based on Bisectors

The mathematical skeleton of a complex form has been essential for a variety of scientific fields and of great interest to many researchers for decades. It is based on several concepts such as the reconstruction of forms and image processing. This paper aims to develop a novel mathematical algorithm to approximate the skeleton of a non-polygonal shape and to compare it to the most used methods. The mathematical technique of skeletonization is used as a reference to validate and compare the proposed method to the most used ones. The crux of the proposed technique is to Cartesianize the shape (polygonize in 2D), then skeletonize it. Moreover, this novel method is grounded upon the construction of bisectors on the simplex of the corresponding Cartesianized shape. Python is used to implement the algorithm proposed and test it on multiple shapes. The comparison of the results generated by the proposed algorithm and the Python predefined function skeletonize() shows that the proposed method is more effective and could be adjusted through the rate of Cartesianization of the target shape. The major contributions of this novel technique include the mitigation of some issues of existing methods, simplification, and optimization of the processing performance mainly in terms of algorithm complexity.

Optimization design method for ultrasonic transducers based on top basic skeleton method and sound pressure distribution in the sound field

Optimization design method for ultrasonic transducers based on top basic skeleton method and sound pressure distribution in the sound field

CORAL REEF TRANSPLANT SUCCESS RATE IN BONETAMBU ISLAND, SPERMONDE ARCHIPELAGO

The world's coral reefs are under threat from human activities through pollution and habitat changes. This condition is influenced by several factors, such as human activities and natural phenomena. Many efforts have been made by humans to overcome or repair coral reef ecosystems that have been damaged. One of the efforts made is to carry out coral transplants. The main goal of coral transplantation is to improve the quality of coral reefs such as increasing live coral cover, biodiversity and uniqueness of coral topography. The stages of the method carried out in the research include mapping the bottom of the waters (sounding); mapping conditions of water cover; transplantation process and monitoring. This study used the spider skeleton method for the transplant media. The results showed that the percentage of live coral cover ranged from 10% - 25% and it was found that rubble cover dominated up to >70%. During the monitoring process, it was found that algae dominated almost all of the transplant media. The high nutrients and the lack of algae-eating fish resulted in a high algae cover found on the surface of the structure and had an impact on the death of the transplanted corals. Efforts are being made to regularly clean algae to increase the survival rate and replant dead coral

Comparative Study of Automated Algorithms for Brain Arteriovenous Malformation Nidus Extent Identification Using 3DRA.

When performing a brain arteriovenous malformation (bAVMs) intervention, computer-assisted analysis of bAVMs can aid clinicians in planning precise therapeutic alternatives. Therefore, we aim to assess currently available methods for bAVMs nidus extent identification over 3DRA. To this end, we establish a unified framework to contrast them over the same dataset, fully automatising the workflows. We retrospectively collected contrast-enhanced 3DRA scans of patients with bAVMs. A segmentation network was used to automatically acquire the brain vessels segmentation for each case. We applied the nidus extent identification algorithms over each of the segmentations, computing overlap measurements against manual nidus delineations. We evaluated the methods over a private dataset with 22 3DRA scans of individuals with bAVMs. The best-performing alternatives resulted in [Formula: see text] and [Formula: see text] dice coefficient values. The mathematical morphology-based approach showed higher robustness through inter-case variability. The skeleton-based approach leverages the skeleton topomorphology characteristics, while being highly sensitive to anatomical variations and the skeletonisation method employed. Overall, nidus extent identification algorithms are also limited by the quality of the raw volume, as the consequent imprecise vessel segmentation will hinder their results. Performance of the available alternatives remains subpar. This analysis allows for a better understanding of the current limitations.

Survey of Noise-Against Techniques for Extracting Stable Skeleton

Stable skeletons are promising as a compact, concise, and efficient descriptor since they can reflect much critical information about the original object, such as topology, connectivity, etc. However, extracting stable skeletons from images is very challenging since most existing skeleton extraction methods, also named skeletonization methods in some literature, are sensitive to noise, which limits the application of the skeletons in the recognition field. Many denoising methods have been proposed to extract stable skeletons to overcome this problem. However, up to now, there are few review papers to conclude these denoising methods and present their pros and cons. Therefore, In this paper, we survey existing denoising techniques for extracting stable skeletons from images. We first categorize these denoising methods, analyze their core idea, and then present their pros and cons for comparison. In addition, we also offer the potential research direction and possibly challenge.

A conveyor belt characterization information extraction method based on YOLOv5 and the skeleton method

To resolve the inaccurate localization of conveyor belt surface damage identification problem and to address the insufficiencies of the methods for extracting surface characterization information, this paper proposes a conveyor belt characterization information extraction method that integrates YOLOv5 deep learning and the skeleton method. By constructing a conveyor belt surface damage recognition model based on the YOLOv5 target detection algorithm, the identification, localization and cropping of the conveyor belt’s surface damage are implemented. After that, edge extraction and surface information extraction are also performed on the damaged parts. Finally, the collected data are analyzed and processed in real time by edge computing equipment to determine the degree of damage of the parts. Finally, intelligent operation of the belt conveyor is achieved with autonomous operations, unattended operations and decision alarms. The experimental results show that the recognition accuracy of YOLOv5 is approximately 93.11%, the speed is approximately 57 frames per second and the error of the data acquired by image processing is between 2% and 10%, which meets the real-time detection requirements of conveyor belt surface damage detection, and assists in the safety management supervision of the belt conveyer.

The Influence of Layer Stacking Method on the Mechanical Properties of Honeycomb Skeleton.

The performance of a multi-layer honeycomb skeleton can be significantly enhanced through tandem connection, while the structure's properties can be tailored by altering the layer stacking method of the honeycomb skeleton. To investigate the impact of layer stacking methods on the mechanical properties of multilayer honeycomb skeletons, 3D printing technology was used to prepare double-layer honeycomb skeleton tandem structures with different dislocation modes in compression testing. A finite element simulation model was established to conduct quasi-static simulation research. Compared to that of a single-layer honeycomb skeleton, the energy absorption of the honeycomb skeleton tandem structure increased. The optimal bearing capacity of the honeycomb skeleton was achieved when the upper and lower layers were precisely aligned. Once dislocation occurred, both the value of average platform stress and energy absorption decreased. Then, the bearing capacity of the honeycomb skeleton tandem structures increased with an enlargement of the dislocation, reaching its maximum at the half-dislocation period. An increase in the partition thickness and stiffness led to a reduction in the dislocation-induced effects on the mechanical properties. The research results can provide theoretical and data support for the engineering application of honeycomb skeleton tandem structures.

Identification of stress transfer in asphalt mixtures based on a reconstruction analysis of the simplified coarse aggregate skeleton

Studies on asphalt mixture skeletons have primarily focused on the statistical skeleton morphology while neglecting internal stress transfer. This study therefore proposed a simplified reconstruction-based coarse aggregate skeleton method for analyzing asphalt mixes using the force chain concept from granular matter mechanics. The stress transfer paths and load transfer efficiencies were investigated for three typical aggregate gradations (AC-13, SMA-13, and OGFC-13) and an analysis of their correlation with uniaxial penetration test results was performed. The AC-13 mixture exhibited the largest uniaxial penetration strength (3.30 kN on average) at the lowest failure displacement (4.1 mm on average). As the characteristics of the simplified skeleton were shown to be related to the number of contact points provided between large aggregate particles, the force chain morphology was closely linked to the mixture gradation. The ratio of strong to weak force chains in the AC-13 mixture was approximately 0.77, whereas those of the SMA-13 and OGFC-13 mixtures were significantly different at 0.90 and 1.13, respectively. The transfer efficiencies of the AC-13, SMA-13, and OGFC-13 specimen skeletons considering morphology alone (i.e., under the no-load condition) were 20.94 %, 62.34 %, and 73.57 %, respectively. However, the correlation between the load transfer efficiency of the coarse aggregate skeleton under a specific load and the observed penetration strength reached 0.9. Thus, the simplified skeleton analysis method proposed herein was shown to satisfactorily retain the skeletal characteristics of the mixes.

Noise-Against Skeleton Extraction Framework and Application on Hand Gesture Recognition

Extracting stable skeletons from noisy images is a challenging problem since the skeletonization method is prone to be affected by inner and border noise. Although many methods have been proposed in the past for increasing the antinoise ability of skeletonization methods, most of them either only overcome border noise or, at the cost of lost topology, degrade the effects of two noises. In this paper, we propose a skeleton extraction framework to enhance the robustness of the existing skeletonization method against both inner and border noise. In our approach, we first use the different scales of Gaussian filters to smooth the input image and obtain multiple representations. Then, binarization and skeletonization were performed to produce a series of binary images and a series of skeletal images. Next, we use our measure on these binary and skeletal images to find the most suitable skeleton. Since our measure considers both the skeleton image changes and binary image changes caused by using a filter, the selected skeleton is sufficiently robust and has all the necessary skeletal branches. The inner noise experiment and border noise experiment are conducted for comparison. From the perspective of the measure of the rate of variation in the skeleton, the proposed framework can reduce the inner noise by approximately 92% and the border noise by approximately 40%. In addition, the experiment on static hand gesture recognition has demonstrated that the introduction of our framework can increase up to 11% mean recognition accuracy.

A novel soft end effector with active palm and fingertips for robotic picking

Designing a type of dexterous soft end effector is the premise to guarantee the working efficiency of a picking robot. For the practical needs of grasping flat objects on the table and unscrewing the bottle cap, we have refined the advantages of traditional soft grippers and designed a novel two-finger soft end effector with active palm and also with two active fingertips. The bending of the fingers actuated by vacuum, the palm and fingertips are driven by motors. Three grasping strategies are developed for the active fingertips so that the gripper can not only grasp flat objects on the tabletop, but also unscrew a cap of a bottle. Further, benefit from the active palm and the high stiffness design method of the finger skeleton, two working modes, pinching mode and grasping mode, have been developed for the gripper. The end effector achieves a maximum grasping force of 32 N in the horizontal direction and 28 N in the vertical direction under the air pressure of −90 kPa. The working angle of the fingers is 0 °–120 °. The size of the object can be grasped ranges from 0 to 200 mm in diameter. Performance verification experiment shows that the active fingertips play an important role in helping the gripper grasp thin objects such as credit cards and printing papers. Grasping experiments on different objects demonstrate that the active fingertips design significantly enhances the flexibility of the gripper and the application it can empower the manipulator.

A Review of Inhibit the Growth of Lithium Dendrite Strategies

Nowadays, the development of electronic technology has driven the development of battery industry. Under the rigid demand for energy storage materials, lithium metal has received a lot of attention due to its excellent energy storage performance, however, the growth of lithium dendrite makes it difficult to recycle. This review introduces the principle of lithium dendrite growth and its negative impact leading to the degradation of battery performance, and then focuses on the methods to inhibit lithium dendrite growth e.g., fabrication of alloyed structure, regulation of solid electrolyte interface (SEI), application of solid electrolyte and recent research progress e.g., nanodiamond additive method, single-atom zinc anion skeleton method, battery self-heating method.

Analysis of the topological motifs of the cellular structure of the tri-spine horseshoe crab (Tachypleus tridentatus) and its associated mechanical properties

Topological motifs in pore architecture can profoundly influence the structural properties of that architecture, such as its mass, porosity, modulus, strength, and surface permeability. Taking the irregular cellular structure of the tri-spine horseshoe crab as a research model, we present a new approach to the quantitative description and analysis of structure-property-function relationships. We employ a robust skeletonization method to construct a curve-skeleton that relies on high-resolution 3D tomographic data. The topological motifs and mechanical properties of the long-range cellular structure were investigated using the Grasshopper plugin and uniaxial compression test to identify the variation gradient. Finite element analysis was conducted for the sub-volumes to obtain the variation in effective modulus along the three principal directions. The results show that the branch length and node distribution density varied from the tip to the base of the sharp corner. These node types formed a low-connectivity network, in which the node types 3-N and 4-N tended to follow the motifs of ideal planar triangle and tetrahedral configurations, respectively, with the highest proportion of inter-branch angles in the angle ranges of 115–120° and 105–110°. In addition, mapping the mechanical gradients to topological properties indicated that narrower profiles with a given branch length gradient, preferred branch orientation, and network connectedness degree are the main factors that affect the mechanical properties. These factors suggest significant potential for designing a controllable, irregularly cellular structure in terms of both morphology and function.

Combined Skeleton and Spatial Rigidification of AIEgens in 2D Covalent Organic Frameworks for Boosted Fluorescence Emission and Sensing of Antibiotics.

AIEgens show relatively weak fluorescence performance owing to the existence of π-π interlayer accumulation, molecular layer planarization, and intramolecular rotation in aggregation-induced emission (AIE) molecules, which limit its application scope. Herein, we put forward a combined skeleton and spatial rigidification method to boost the fluorescence emission efficiency of AIEgens. As a proof-of-concept experiment, two highly fluorescent covalent organic frameworks (COFs) were designed and constructed by the Knoevenagel condensation reaction. The experimental results show that the combined skeleton and spatial rigidification endowed excellent fluorescence emission for the resulting F-COF-2 by destruction of the π-π interlayer accumulation, interference of the molecular layer planarization, and restriction of the intramolecular rotation of the AIEgen unit. F-COF-2 displayed highly sensitive and selective NFT and NZF detection. Particularly, the Ksv value and limit of detection of F-COF-2 toward NFT were estimated to be 9.12 × 105 M-1 and 3.35 ppb, respectively, which surpassed all the reported crystalline porous fluorescent materials. The mechanism study proved that its outstanding fluorescence detection property was ascribed to the formation of a nonfluorescent complex induced by hydrogen bond interactions and electron transfer between F-COF-2 and NFT and NZF. This work not only proposes a combined skeleton and spatial rigidification strategy to improve the fluorescence efficiency of AIE molecules but also develops a sensor with high fluorescence efficiency, high chemical stability, and highly efficient detection of antibiotics.

Extracting Skeletons of Two-Manifold Triangular Mesh Surface for Planning Skeleton-Guided Five-Axis Surface Inspection Path

Abstract Skeleton curve-guided five-axis sweep scanning is an emerging surface inspection technique with marvelous inspection efficiency. The precondition of applying such a technique to a complex surface is to extract the skeleton curves of a two-manifold surface associated with partitioning the surface into compact surfaces. However, the work on extracting skeletons of two-manifold triangle mesh surfaces is scarce, as existing skeletonization methods mostly focus on either 2D planar shapes or 3D solid shapes. In this paper, we present a new approach to extract the skeletons on a two-manifold triangle mesh with boundaries. The skeletons are formed in the most intuitive manner of wave front propagation which is based on computing the initial value problem of geodesics on a triangle mesh. The step-length of wave front propagation is adaptively controlled to guarantee the appropriate density of skeleton points for having good connectivity. Experiments show that, as a direct application of the proposed skeleton generation and its associated surface partitioning result, the five-axis sweep scanning path of complex free-form surface can be generated conveniently. Experimental results also validate that the computation of our proposed approach is simpler and faster than the state-of-the-art geodesic Voronoi diagram (GVD) method when most of the two-manifold triangle mesh is nearly planar. Additionally, the augmented parameter used for tracing the wave front’s geodesic propagation information is helpful for the skeleton-based surface partition, which is necessary for skeleton curve-guided five-axis sweep scanning.

Dynamic Hand Gesture Recognition for Smart Lifecare Routines via K-Ary Tree Hashing Classifier

In the past few years, home appliances have been influenced by the latest technologies and changes in consumer trends. One of the most desired gadgets of this time is a universal remote control for gestures. Hand gestures are the best way to control home appliances. This paper presents a novel method of recognizing hand gestures for smart home appliances using imaging sensors. The proposed model is divided into six steps. First, preprocessing is done to de-noise the video frames and resize each frame to a specific dimension. Second, the hand is detected using a single shot detector-based convolution neural network (SSD-CNN) model. Third, landmarks are localized on the hand using the skeleton method. Fourth, features are extracted based on point-based trajectories, frame differencing, orientation histograms, and 3D point clouds. Fifth, features are optimized using fuzzy logic, and last, the H-Hash classifier is used for the classification of hand gestures. The system is tested on two benchmark datasets, namely, the IPN hand dataset and Jester dataset. The recognition accuracy on the IPN hand dataset is 88.46% and on Jester datasets is 87.69%. Users can control their smart home appliances, such as television, radio, air conditioner, and vacuum cleaner, using the proposed system.