The human skeleton establishment aims to provide accurate localization information of the human body from RGB images and establish a complete human skeleton for many applications, such as action recognition, video surveillance, and human-computer interaction. Considering the inherent human body structure, many recent methods group the relevant body parts and utilize the deep convolutional network to learn the visual context from the part groups. However, the grouping approaches used in these methods heavily rely on prior knowledge of the human body shape but lose important relationships between parts. In this paper, we introduce the Accurate Part Grouping Network (Accurate-PGNet), a novel network for hierarchically grouping body parts in a data-driven manner. In contrast to the previous methods, we use neural architecture search (NAS) to optimize the architecture of Accurate-PGNet and properly group the body parts. The part grouping respects the diverse visual patterns of parts, producing groups containing different body parts. From each group, we learn the visual feature map. It helps to capture the correlation between parts and predict their locations. The feature maps of the part groups are merged hierarchically to capture the higher-order context of parts in larger groups. We extensively evaluated our method on the challenging benchmarks, demonstrating that Accurate-PGNet effectively helps to achieve state-of-the-art results.

Learning 3-D structures from incomplete point clouds with extreme sparsity and random distributions is a challenge since it is difficult to infer topological connectivity and structural details from fragmentary representations. Missing large portions of informative structures further aggravates this problem. To overcome this, a novel graph convolutional network (GCN) called dynamic and structure-aware NETwork (DSANet) is presented in this article. This framework is formulated based on a pyramidic auto-encoder (AE) architecture to address accurate structure reconstruction on the sparse and incomplete point clouds. A PointNet-like neural network is applied as the encoder to efficiently aggregate the global representations of coarse point clouds. On the decoder side, we design a dynamic graph learning module with a structure-aware attention (SAA) to take advantage of the topology relationships maintained in the dynamic latent graph. Relying on gradually unfolding the extracted representation into a sequence of graphs, DSANet is able to reconstruct complicated point clouds with rich and descriptive details. To associate analogous structure awareness with semantic estimation, we further propose a mechanism, called structure similarity assessment (SSA). This method allows our model to surmise semantic homogeneity in an unsupervised manner. Finally, we optimize the proposed model by minimizing a new distortion-aware objective end-to-end. Extensive qualitative and quantitative experiments demonstrate the impressive performance of our model in reconstructing unbroken 3-D shapes from deficient point clouds and preserving semantic relationships among different regional structures.

In the task of dynamic human pose estimation (dynamic HPE), the temporal relationships between human body parts should be captured comprehensively to understand the dynamic human motions, where the correlated motion information eventually helps to recognize body parts. The popular methods are successful in terms of utilizing long-term motion information captured by low-speed cameras. Yet they neglect the underlying intermediate motions between captured frames, which comprise the temporal-interim poses lost in the video. In this article, we introduce a novel framework, temporal-interim pose synthesis and distillation, to produce and leverage the intermediate motion information for dynamic motion establishment. The pose synthesis yields the visual feature maps of the intermediate poses, which appear between the existing video frames. It allows the synthesized and current poses to form richer motion patterns. Next, the pose distillation divides the body parts into several groups, where it learns the specific part-wise relationship within each group. It degrades the complexity of learning useful part-wise relationships from rich motion patterns and extracts more detailed motion information for fine-grained part groups. We extensively evaluate our method on challenging datasets for dynamic pose estimation, achieving state-of-the-artresults.

Abstract Reconstructing and animating digital avatars with free views from monocular videos have been an interesting research task in the computer vision field for a long time. Recently, some methods have introduced a novel category method of leveraging the neural radiance field to represent the human body in a canonical space with the help of the SMPL model. With the deformation of the points from an observation space into a canonical space, the human appearance can be learned in various poses and viewpoints. However, previous methods highly rely on pose-dependent representation learned from frame-independent optimization and ignore the temporal contexts across the continuous motion video, causing a bad influence on the dynamic appearance texture generation. To overcome these problems, we propose a novel free-viewpoint rendering framework, TMIHuman. It aims at introducing temporal information into NeRF-based rendering and distilling task-relevant information from complex pixel-wise representations. To be specific, we build a temporal fusion encoder that imports timestamps into the learning of non-rigid deformation and fuses the visual features of other frames into human representation. Then, we propose to disentangle the fused features and extract useful visual cues via mutual information objectives. We have extensively evaluated our method and achieved state-of-the-art performance on different public datasets.

Point clouds are fundamental in the representation of 3D objects. However, they can also be highly unstructured and irregular. This makes it difficult to directly extend 2D generative models to three-dimensional space. In this article, we cast the problem of point cloud generation as a topological representation learning problem. In order to capture the representative features of 3D shapes in the latent space, we propose a hierarchical mixture model that integrates self-attention with an inference tree structure for constructing a point cloud generator. Based on this, we design a novel Generative Adversarial Network (GAN) architecture that is capable of generating recognizable point clouds in an unsupervised manner. The proposed adversarial framework (SG-GAN) relies on self-attention mechanism and Graph Convolution Network (GCN) to hierarchically infer the latent topology of 3D shapes. Embedding and transferring the global topology information in a tree framework allows our model to capture and enhance the structural connectivity. Furthermore, the proposed architecture endows our model with partially generating 3D structures. Finally, we propose two gradient penalty methods to stabilize the training of SG-GAN and overcome the possible mode collapse of GAN networks. To demonstrate the performance of our model, we present both quantitative and qualitative evaluations and show that SG-GAN is more efficient in training and it exceeds the state-of-the-art in 3D point cloud generation.

Visual query of spatiotemporal data is becoming an increasingly important function in visual analytics applications. Various works have been presented for querying large spatiotemporal data in real time. However, the real-time query of spatiotemporal data distribution is still an open challenge. As spatiotemporal data become larger, methods of aggregation, storage and querying become critical. We propose a new visual query system that creates a low-memory storage component and provides real-time visual interactions of spatiotemporal data. We first present a peak-based kernel density estimation method to produce the data distribution for the spatiotemporal data. Then a novel density dictionary learning approach is proposed to compress temporal density maps and accelerate the query calculation. Moreover, various intuitive query interactions are presented to interactively gain patterns. The experimental results obtained on three datasets demonstrate that the presented system offers an effective query for visual analytics of spatiotemporal data.

Point clouds are the most general data representations of real and abstract objects, and have a wide variety of applications in many science and engineering fields. Point clouds also provide the most scalable multi-resolution composition for geometric structures. Although point cloud learning has shown remarkable results in shape estimation and semantic segmentation, the unsupervised generation of 3D object parts still pose significant challenges in the 3D shape understanding problem. We address this problem by proposing a novel Generative Adversarial Network (GAN), named HSGAN, or Hierarchical Self-Attention GAN, with remarkable properties for 3D shape generation. Our generative model takes a random code and hierarchically transforms it into a representation graph by incorporating both Graph Convolution Network (GCN) and self-attention. With embedding the global graph topology in shape generation, the proposed model takes advantage of the latent topological information to fully construct the geometry of 3D object shapes. Different from the existing generative pipelines, our deep learning architecture articulates three significant properties HSGAN effectively deploys the compact latent topology information as a graph representation in the generative learning process and generates realistic point clouds, HSGAN avoids multiple discriminator updates per generator update, and HSGAN preserves the most dominant geometric structures of 3D shapes in the same hierarchical sampling process. We demonstrate the performance of our new approach with both quantitative and qualitative evaluations. We further present a new adversarial loss to maintain the training stability and overcome the potential mode collapse of traditional GANs. Finally, we explore the use of HSGAN as a plug-and-play decoder in the auto-encoding architecture.

Connectivity analysis between the components of large evolving systems can reveal significant patterns of interaction. The systems can be simulated by topological graph structures. However, such analysis becomes challenging on large and complex graphs. Tasks such as comparing, searching, and summarizing structures, are difficult due to the enormous number of calculations required. For time-varying graphs, the temporal dimension even intensifies the difficulty. In this article, we propose to reduce the complexity of analysis by focusing on subgraphs that are induced by closely related entities. To summarize the diverse structures of subgraphs, we build a supervised layout-based classification model. The main premise is that the graph structures can induce a unique appearance of the layout. In contrast to traditional graph theory-based and contemporary neural network-based methods of graph classification, our approach generates low costs and there is no need to learn informative graph representations. Combined with temporally stable visualizations, we can also facilitate the understanding of sub-structures and the tracking of graph evolution. The method is evaluated on two real-world datasets. The results show that our system is highly effective in carrying out visual-based analytics of large graphs.

Abstract In urban analysis, it is desirable to find regions where a primary socio‐economic activity dominates as a key endeavour. This can be accomplished by aggregating neighbouring locations where similar activities take place. However, people move and their activities change over time. Furthermore, the boundaries of regions are not stationary. Thus, it is challenging to update region divisions and track their evolution. Geo‐textual data embody geographical information and activity descriptions. We obtain changes in regional boundaries by iteratively applying a community detection process to a sequence of latent graphs that are constructed from geo‐textual data. Region characteristics are interpreted by topics learned by the latent Dirichlet allocation model. We also propose a matching algorithm to expose region transformations between different timestamps. Interesting patterns of evolution emerge after clustering the migration trajectories of region centroids. In our visual system, users can explore the evolution of regions through animations and linked snapshots. To facilitate visual comparisons, we represent regions by hexagonal tiling that better construct arbitrary regional shapes. The effectiveness of our method is evaluated on two case studies using real‐world datasets, and a user study shows that our visual analytics system is highly effective in performing studies on such regional maps.

Abstract Feature matching is the most basic and pervasive problem in computer vision and it has become a primary component in big data analytics. Many tools have been developed for extracting and matching features in video streams and image frames. However, one of the most basic tools, that is, a tool for simply visualizing matched features for the comparison and evaluation of computer vision algorithms is not generally available, especially when dealing with a large number of matching lines. We introduce VisFM, an integrated visual analysis system for comprehending and exploring image feature matchings. VisFM presents a matching view with an intuitive line bundling to provide useful insights regarding the quality of matched features. VisFM is capable of showing a summarization of the features and matchings through group view to assist domain experts in observing the feature matching patterns from multiple perspectives. VisFM incorporates a series of interactions for exploring the feature data. We demonstrate the visual efficacy of VisFM by applying it to three scenarios. An informal expert feedback, conducted by our collaborator in computer vision, demonstrates how VisFM can be used for comparing and analysing feature matchings when the goal is to improve an image retrieval algorithm.

Interactive visualization of streaming points for real-time scatterplots and linear blending of correlation patterns is increasingly becoming the dominant mode of visual analytics for both big data and streaming data from active sensors and broadcasting media. To better visualize and interact with inter-stream patterns, it is generally necessary to smooth out gaps or distortions in the streaming data. Previous approaches either animate the points directly or present a sampled static heat-map. We propose a new approach, called StreamMap, to smoothly blend high-density streaming points and create a visual flow that emphasizes the density pattern distributions. In essence, we present three new contributions for the visualization of high-density streaming points. The first contribution is a density-based method called super kernel density estimation that aggregates streaming points using an adaptive kernel to solve the overlapping problem. The second contribution is a robust density morphing algorithm that generates several smooth intermediate frames for a given pair of frames. The third contribution is a trend representation design that can help convey the flow directions of the streaming points. The experimental results on three datasets demonstrate the effectiveness of StreamMap when dynamic visualization and visual analysis of trend patterns on streaming points are required.

A color design process for fabric images can resort to a solution of a color transfer problem based on given color themes. Usually, the color transfer process contains an image segmentation phase and an image construction phase. In this paper, a novel color transfer method for fabric images is proposed. Compared with classical color transfer methods, the new method has the following three main innovations. First, the new method, in its image segmentation phase, follows an assumption that a fabric image can be decomposed into cartoon and texture components, which means the new color transfer method, in its image segmentation, phase incorporates an image decomposition process. The advantage of the innovation is that the cartoon component is more suitable than the original image to be used to partition the fabric image. Second, the new color transfer method can generate more vivid color transfer results since the above texture component is used to describe yarn texture details in the image construction phase. Third, the total generalized variation (TGV) regularizer is used to further improve the performance of image decomposition. Here, the TGV regularizer is good at estimating the weak lightness variation of the cartoon component with the CIELab color scheme. In addition, by using the augmented Lagrange multiplier method, we derive an efficient algorithm to search for the solutions to the proposed color transfer problem. Numerical results demonstrate that the proposed color transfer method can generate better results for fabric images.

In this paper, visual attention spreading is formulated as a nonlocal diffusion equation. Different from other diffusion-based methods, a nonlocal diffusion tensor is introduced to consider both the diffusion strength and the diffusion direction. With the help of diffusion tensor, along with the principle direction, the diffusion has been suppressed to preserve the dissimilarity between the foreground and background, while in other directions, the diffusion has been boosted to combine the similar regions and highlight the salient object as a whole. Through a two-stages diffusion, the final saliency maps are obtained. Extensive quantitative or visual comparisons are performed on three widely used benchmark datasets, i.e. MSRA-ASD, MSRA-B and PASCAL-1500 datasets. Experimental results demonstrate the superior performance of our method.

Bag-of-visual-words (BOVW)-based image representation has received intense attention in recent years and has improved content-based image retrieval (CBIR) significantly. BOVW does not consider the spatial correlation between visual words in natural images and thus biases the generated visual words toward noise when the corresponding visual features are not stable. This article outlines the construction of a visual word co-occurrence matrix by exploring visual word co-occurrence extracted from small affine-invariant regions in a large collection of natural images. Based on this co-occurrence matrix, we first present a novel high-order predictor to accelerate the generation of spatially correlated visual words and a penalty tree (PTree) to continue generating the words after the prediction. Subsequently, we propose two methods of co-occurrence weighting similarity measure for image ranking: Co-Cosine and Co-TFIDF. These two new schemes down-weight the contributions of the words that are less discriminative because of frequent co-occurrences with other words. We conduct experiments on Oxford and Paris Building datasets, in which the ImageNet dataset is used to implement a large-scale evaluation. Cross-dataset evaluations between the Oxford and Paris datasets and Oxford and Holidays datasets are also provided. Thorough experimental results suggest that our method outperforms the state of the art without adding much additional cost to the BOVW model.

In this paper, we propose a new fuzzy-based variational model that efficiently computes partitioning of speckled images, such as images obtained from Synthetic Aperture Radar (SAR). The model is derived by using the so-called maximizing a posteriori (MAP) estimation method. The novelties of the model are: (1) the Gamma distribution rather than the classical Gaussian distribution is used to model the gray intensities in each homogeneous region of the images (Gamma distribution function is better suited for speckled images); (2) an adaptive weighted regularization term with respect to a fuzzy membership function is designed to protect the segmentation results from degeneration (being over-smoothed). Compared with the classical total variation (TV) regularizer, the proposed regularization term has a sparser property. In addition, a new alternative direction iteration algorithm is proposed to solve the model. The algorithm is efficient since it integrates the split Bregman method and the Chambolle's projection method. Numerical examples are given to verify the efficiency of our model.

Visualization of high density streaming points has become a challenge in information exploration. In this paper, we present a new pipeline for the interactive visualization of large points set. The pipeline is based on the idea that heat-map can overcome the overlapping problem in visualization of high density streaming points. Thus, we firstly define a regular streaming format for large point set which can be updated or changed continually. Based on streaming points, we use kernel density estimation to estimate the point distribution and visualize the density image. Perceptive and interactive features are also considered in our visualization. To our knowledge, our pipeline is the first work that focuses on perceptive visualization of high density streaming points. The main step of our pipeline is accelerated via GPU rendering in order to make scene of real-time interaction in visualization. We demonstrate the visual effectiveness of our pipeline on a geographical dataset of high-density streaming points.

Visual analytics of increasingly large data sets has become a challenge for traditional in-memory and off-line algorithms as well as in the cognitive process of understanding features at various scales of resolution. In this paper, we attempt a new web-based framework for the dynamic visualization of large data. The framework is based on the idea that no physical device can ever catch up to the analytical demand and the physical requirements of large data. Thus, we adopt a data streaming generator model that serializes the original data into multiple streams of data that can be contained on current hardware. Thus, the scalability of the visual analytics of large data is inherent in the streaming architecture supported by our platform. The platform is based on the traditional server-client model. However, the platform is enhanced by effective analytical methods that operate on data streams, such as binned points and bundling lines that reduce and enhance large streams of data for effective interactive visualization. We demonstrate the effectiveness of our framework on different types of large datasets.

In this brief, we show that minimizing nearest neighbor classification error (MNNE) is a favorable criterion for supervised linear dimension reduction (SLDR). We prove that MNNE is better than maximizing mutual information in the sense of being a proxy of the Bayes optimal criterion. Based on kernel density estimation, we derive a nonparametric algorithm for MNNE. Experiments on benchmark data sets show the superiority of MNNE over existing nonparametric SLDR methods.

The efficient and robust extraction of invariant patterns from an image is a long-standing problem in computer vision. Invariant structures are often related to repetitive or near-repetitive patterns. The perception of repetitive patterns in an image is strongly linked to the visual interpretation and composition of textures. Repetitive patterns are products of both repetitive structures as well as repetitive reflections or color patterns. In other words, patterns that exhibit near-stationary behavior provide rich information about objects, their shapes, and their texture in an image. In this paper, we propose a new algorithm for repetitive pattern detection and grouping. The algorithm follows the classical region growing image segmentation scheme. It utilizes a mean-shift-like dynamic to group local image patches into clusters. It exploits a continuous joint alignment to: 1) match similar patches, and 2) refine the subspace grouping. We also propose an algorithm for inferring the composition structure of the repetitive patterns. The inference algorithm constructs a data-driven structural completion field, which merges the detected repetitive patterns into specific global geometric structures. The result of higher level grouping for image patterns can be used to infer the geometry of objects and estimate the general layout of a crowded scene.

The appearance of woven fabrics is intrinsically determined by the geometric details of their meso/micro scale structure. In this paper, we propose a multiscale representation and tessellation approach for woven fabrics. We extend the Displaced Subdivision Surface (DSS) to a representation named Interlaced/Intertwisted Displacement Subdivision Surface (IDSS). IDSS maps the geometric detail, scale by scale, onto a ternary interpolatory subdivision surface that is approximated by Bezier patches. This approach is designed for woven fabric rendering on DX11 GPUs. We introduce the Woven Patch, a structure based on DirectX’s new primitive, patch, to describe an area of a woven fabric so that it can be easily implemented in the graphics pipeline using a hull shader, a tessellator and a domain shader. We can render a woven piece of fabric at 25 frames per second on a low-performance NVIDIA 8400 MG mobile GPU. This allows for large-scale representations of woven fabrics that maintain the geometric variances of real yarn and fiber.

The efficient and robust extraction of invariant patterns from an image is a long-standing problem in computer vision. Invariant structures are often related to repetitive or near-repetitive patterns. The perception of repetitive patterns in an image is strongly linked to the visual interpretation and composition of textures. Repetitive patterns are products of both repetitive structures as well as repetitive reflections or color patterns. In other words, patterns that exhibit near-stationary behavior provide a rich information about objects, their shapes, and their texture in an image. In this paper, we propose a new algorithm for repetitive pattern detection and grouping. The algorithm follows the classical region growing image segmentation scheme. It utilizes a mean-shift-like dynamics to group local image patches into clusters. It exploits a continuous joint alignment to (a) match similar patches and (b) refine the subspace grouping. The result of higher-level grouping for image patterns can be used to infer the geometry of object surfaces and estimate the general layout of a crowded scene.

Visual patterns generated by color patches, texture regions, and repetitive textons in an image can be organized into higher-level structural forms such as geometric shapes, arrays, and partition groups. Understanding the information content formed by these visual pattern compositions is important both from a theoretical point of view as well as in the robust implementation of many image processing applications. In this paper we propose a new method for building pattern compositions and inferring the high-level structural forms over near regular textures. We exploit the shape geometry of repeated patterns to interpret pairwise connections between patterns and generate the abstract structural form by unifying the local connections. The inferred structure can reflect the organization of multiple repeated patterns and can be used in the classification of texture structures.

The articles in this special section contain selected papers from the 2010 ACM Virtual Reality Software and Technology Symposium.

Recently, in the context of IEEE 802.11b/g network protocols, Wi-Fi radio channels been proposed to estimate the location of a smart mobile device. We can locate Wi-Fi-enabled devices by applying location-sensing techniques. However, the positioning accuracy depends greatly on the Wi-Fi signal coverage. The positioning accuracy due to poor Wi-Fi signal coverage has not been investigated systematically in the current research on Wi-Fi location awareness. Our previous work provide a location threshold of 1.82m on average. However, when a person enters in a poor Wi-Fi coverage region, the positioning accuracy drops dramatically. In this paper, we extend our previous work and create a fuzzy color map to visualize the distribution of Wi-Fi signal: red represents strong signals and blue represents weak signals. Then we make use of the proposed map by selecting the best candidates of AP to increase the positioning accuracy in the poor Wi-Fi coverage region. Our experiment result shows that we can reduce the distance error significantly by 25% in a poor Wi-Fi coverage environment and locate a person within 1.75m in average. The proposed method leads to substantially more accurate and robust localization system.

Generation of a large-scale city requires a significant amount of manual work and computation to process massive location information and model building geometry with multi-level of details. Normally, an urban city is heavily built-up with different architectural building patterns across extensively and topographically varied landscapes. In this paper, we introduce Footprint-Profile Sweep Surfaces (FPSS), a flexible and computationally efficient approach for realtime generation and rendering of massive urban buildings in a heavily built-up city. A solid constituting an urban building is represented as an instance of FPSS and is generated by sweeping a footprint along a profile with specific parameters. We present two forms of FPSS: super FPSS to address the shapes from architecture design and poly FPSS to address the shapes from imported GIS data. We make use of hardware tessellation to allow dynamic LOD according to view distance. A special scaling-translation-rotation displacement performed on the simplified profile is proposed to support detail generation. Experimental results show that realtime performance can be achieved using our approach to generate varied styles of urban buildings. Even inexperienced users are able to generate a building group quickly in their own style based on FPSS.

Continuous collision detection improves the computation of the contact information for interacting objects in dynamic virtual environments. The computation cost is relatively high in the phase of the elementary test processing. In virtual environments, such as crowds in large urban models, there is a large portion of feature pairs that do not collide but the computation is relatively of high cost. In this paper, we propose a robust approach for solving the scalability of the collision detection problem by applying four distinct phases. First, k-DOPs are used for culling non-proximal triangles. Second, the feature assignment scheme is used for minimizing the number of potentially colliding feature pairs. Third, an intrinsic filter is employed for filtering non-coplanar feature pairs. Forth, we use a direct method for computing the contact time that is more efficient than the numerical Interval Newton method. We have implemented our system and have compared its performance with the most recently developed approaches. Six benchmarks were evaluated and the complexity of the models was up to 1.5M triangles. The experimental results show that our method improves the performance for the elementary tests.

Woven fabrics are widely used in clothing because of their parallel and interlaced properties, which are formed by weaving. Creating a weaving pattern, especially hand weaving for interlacing yarns is a cumbersome task in the textile industry. In this paper, we propose two kinds of playing for creating weaving patterns on multi-input devices: the tie-up plan and the lift plan. Discrete notes on the treble staff are translated into signatures of treadling sequences and discrete notes in the bass staff are translated into signatures of theadling sequences. Artists can use their right hand to compose a treadling sequence for weft yarns and their left hand to play a threading sequence for warp yarns. The treadling and threading sequences become the notes on the full gamut of shafts and treadles. Our result shows that we are able to compose a family of weaving patterns in a similar way to playing the piano in a short time.

Localization systems for indoor areas have recently been suggested that make use of existing wireless local area network (WLAN) infrastructure and location fingerprinting approach. However, most existing research work ignores channel interference between wireless infrastructures and this could affect accurate and precise positioning. A better understanding of the properties of channel interference could assist in improving the positioning accuracy while saving significant amounts of resources in the location-aware infrastructure. This paper investigates to what extent the positioning accuracy is affected by channel interference between access points. Two sets of experiments compare how the positioning accuracy is affected in three different channel assignment schemes: ad-hoc, sequential, and orthogonal data is analyzed to understand what features of channel interference affect positioning accuracy. The results show that choosing an appropriate channel assignment scheme could make localization 10% more accurate and reduce the number of access points that are required by 15%. The experimental analysis also indicates that the channel interference usually obeys a right-skewed distribution and positioning accuracy is heavily dependent on channel interference between access points (APs).

While localization systems for indoor areas using the existing wireless local area network (WLAN) infrastructure have recently been proposed, wireless LAN localization approaches suffer from a number of significant drawbacks. To begin with, there is inaccurate position tracking due to the orientation of the mobile device and signal fluctuation. In this paper, we apply an orientation filter and a Newton Trust Region (TR) algorithm to eliminate the noisy location estimation. We implement the localization algorithm on the Nexus One which is a Wi-Fi enabled device with a digital compass. The average error distance is only 1.82m. We achieve 90% precision within 2.45m. The proposed method leads to substantially more accurate and robust localization system.

Localization systems for indoor areas using the existing wireless local area network (WLAN) infrastructure have been suggested recently. However, the current systems are not satisfactory. Common localization approaches suffer from inaccurate position tracking due to signal fluctuations in the wireless LAN. Newton trust-region method makes use of the convergence factor of a trajectory to eliminate the noise from the signal strength. In this paper, we apply a Newton trust-region (TR) algorithm to trajectory estimation based on the traditional location fingerprinting (localization) approach. Newton trust-region method optimize the location fingerprinting approach iteratively because each point in a trajectory normally falls into a region and have the same convergence in direction. Our experimental analysis shows that Newton trust-region method enhances the traditional localization approach with 15% fewer access points and is 20% more effective to achieve accurate localization. The proposed Newton trust-region method leads to substantially more accurate and robust localization system.

Wireless tracking analysis is useful for deploying the efficient indoor positioning system. Location fingerprinting (LF) method uses a training dataset of Wi-Fi received signal strength (RSS) at different location to track the target. Fuzzy logic modeling can be applied to evaluate the behavior of wireless received signal strength (RSS). Previous analytical models based on LF are not sufficient for modeling spatial factors of wireless coverage. Spatial analytical model is useful for analysis of how the wireless infrastructure affecting the accuracy of positioning. The main concept of fuzzy logic is to reflect the reality of our world of experience, which is uncertain and fuzzy. In this paper, we develop a multilayer fuzzy modeling for the wireless coverage in the huge and open area. Large scale site surveying has been used to collect RSS in 9.34 hectare campus area. The color fuzzy model allows us to visualize the spatial distribution of wireless RSS. Base on the fuzzy analytical model, we analyze the effect of existence of human's presence and large obstacle, the accuracy and efficiency of tracking system.

Abstract Collision events between 3D objects in motion in computer animations or simulations are difficult to detect due to the difficulty of accurately sampling the motion paths of objects in space and time. One approach to this problem has been continuous collision detection but because the current approaches process potentially interacting primitive pairs (PIPPs) redundantly. This is time‐expensive, especially where there are a large number of PIPPs. In this paper we propose a novel collision detection process that more accurately and robustly detects collisions on simulated meshed deformable surfaces. We embed a new layer, primitive filtering layer (PFL), to extract PIPPs. This has two results. It reduces the number of PIPPs significantly and it means that each interacting primitive pair is processed just one time. Experimental results show that this approach achieves interactive rates for complex deformable surfaces with large contact regions. This is especially practical for cloth dynamics. Our method is efficient, accurate, reliable, and robust even in the presence of objects with sharp features. We also present techniques to implement the method on programmable graphics processing units (GPUs). Copyright © 2007 John Wiley & Sons, Ltd.

Abstract Research on motion retargeting and synthesis for character animation has been mostly focused on character scale variations. In our recent work we have addressed the motion retargeting problem for characters with slightly different topologies. In this paper we present a new method for retargeting captured motion data to an enhanced character skeleton having a topology that is different from that of the original captured motion. The new topology could include altered hierarchical structures and scaled segments. In order to solve this problem, we propose a framework based on the concept of a motion control net (MCN). This is an external structure analogous to the convex hull of a set of control points defining a parametric curve or a surface patch. The MCN encapsulates the motion characteristics of the character. Retargeting is achieved as a generalized inverse kinematics problem using an external MCN. The retargeting solution requires the dynamic modification of the MCN structure. This also allows US to interactively edit the MCN and modify the conditions for the motion analysis. The new method can automatically synthesize new segment information and, by combining the segment motion into the MCN domain with a suitable displacement of control points embedded in the original motion capture sensor data, it can also generate realistic new motions that resemble the motion patterns in the original data. Copyright © 2006 John Wiley & Sons, Ltd.

Abstract In this paper, we present a new segmentation solution for extracting motion patterns from motion capture data by searching for critical keyposes in the motion sequence. A rank is established for critical keyposes that identifies the significance of the directional change in motion data. The method is based on entropy metrics, specifically the mutual information measure. Displacement histograms between frames are evaluated and the mutual information metric is employed in order to calculate the inter‐frame dependency. The most significant keypose identifies the largest directional change in the motion data. This will have the lowest mutual information level from all the candidate keyposes. Less significant keyposes are then listed with higher mutual information levels. The results show that the method has higher sensitivity in the directional change than methods based on the magnitude of the velocity alone. This method is intended to provide a summary of a motion clip by ranked keyposes, which is highly useful in motion browsing and motion retrieve database system. Copyright © 2005 John Wiley & Sons, Ltd.

Abstract An intrinsic collision detection unit (ICDU) forms the bottom‐most layer of a collision detection pipeline. The ICDU performs collision detection and computes collision information for primitive feature pairs of objects in a 3D dynamic environment. A significant amount of time can be spent by the ICDU during the collision detection process. In this paper, we extend the ICDU framework to take advantages of the computational power of programmable graphics processors (GPUs). Some components of the ICDU framework consist of time demanding and fine‐grained tasks that can be implemented on GPUs. By employing the framework, collision information can be computed accurately, robustly, and efficiently. Experimental results show that the proposed method greatly improves the performance of the ICDU. A collection buffer is proposed for the future enhancement of GPU‐based collision detectors. Copyright © 2005 John Wiley & Sons, Ltd.

In this paper, we introduce new techniques that enhance the computational performance for the interactions between sharp objects and deformable surfaces. The new formulation is based on a time-domain predictor-corrector model. For this purpose, we define a new kind of (pi, beta, I)-surface. The partitioning of a deformable surface into a finite set of (pi, beta, I)-surfaces allows us to prune a large number of noncolliding feature pairs. This leads to a significant performance improvement in the collision detection process. The intrinsic collision detection is performed in the time domain. Although it is more expensive compared to the static interference test, it avoids portions of the surfaces passing through each other in a single time step. In order to resolve all the possible collision events at a given time, a penetration-free motion space is constructed for each colliding particle. By keeping the velocity of each particle inside the motion space, we guarantee that the current colliding feature pairs will not penetrate each other in the subsequent motion. A static analysis approach is adopted to handle friction by considering the forces acting on the particles and their velocities. In our formulation, we further reduce the computational complexity by eliminating the need to compute repulsive forces.

We propose a positional constraint method to solve the multi-layered deformable surface problem based on a master-slave scheme. This allows two or more deformable surfaces to be attached together in any orientation relative to each other for the purpose of modeling cloth attachments and multi-layered clothing. The method does not require the mesh resolution of the deformable surfaces to be the same or the matching of anchor points between layers. After the attachment process, the surfaces are treated as a multi-layered surface. However, this surface contains non-manifold features. We introduce a technique for preventing self-intersection of the non-manifold features. We demonstrate the stability of this method by performing several experiments with high surface complexity and a large number of colliding feature pairs. Interactive rates can easily be achieved for multilayered surfaces with an appropriate discretization level of triangles.

Modeling the natural interaction of cloth and garments with objects in a 3D environment is currently one of the most computationally demanding tasks. These highly deformable materials are subject to a very large number of contact points in the proximity of other moving objects. Furthermore, cloth objects often fold, roll, and drape within themselves, generating a large number of self-collision areas. The interactive requirements of 3D games and physically driven virtual environments make the cloth collisions and self-collision computations more challenging. By exploiting mathematically well-defined smoothness conditions over smaller patches of deformable surfaces and resorting to image-based collision detection tests, we developed an efficient collision detection method that achieves interactive rates while tracking self-interactions in highly deformable surfaces consisting of a large number of elements. The method makes use of a novel technique for dynamically generating a hierarchy of cloth bounding boxes in order to perform object-level culling and image-based intersection tests using conventional graphics hardware support. An efficient backward voxel-based AABB hierarchy method is proposed to handle deformable surfaces which are highly compressed.

Most collision detection methods developed so far are based on geometrical object-space interference tests. While this remains the basic mode of investigation for geometric algorithms, the requirements for interactive rates and complex geometry predominate in commercial applications. In this article, we propose a new mode of collision detection based on an image-space approach. This approach breaks the object-space collision detection bottleneck by distributing the computational load onto the hardware graphics pipeline. The image-space approach, in conjunction with efficient bounding-box strategies in the object-space, has the potential to handle complex object interactions at interactive rates.

A novel algorithm for computing RSIC intersection curves of two surfaces of revolution is presented; it subdivides a surface of revolution into a collection of coaxial spherical strips, by subdividing its generatrix into a collection of C/sup 0/ and C/sup 1/ coaxial circular arcs correspondingly. So the intersection problem of two surfaces of revolution is reduced to an intersection problem of two spherical stripes, and RSIC is approximated as a piecewise C/sup 0/ and C/sup 1/ circular, which is quite convenient for some CAD applications. Cylindrical bounding shell is used for computing valid intersection interval VII to avoid the unnecessary intersection computation efficiently. Finally, we give a simple algorithm of tracing RSIC for classification and a unified rational Bezier representation to RSIC.

The natural behavior of garments and textile materials in the presence of changing object states is potentially the most computationally demanding task in a dynamic 3D virtual environment. Cloth materials are highly deformable inducing a very large number of contact points or regions with other objects. In a natural environment, cloth objects often interact with themselves generating a large number of self-collisions areas. The interactive requirements of 3D games and physically driven virtual environments make the cloth collisions and self-collisions computations more challenging. By exploiting mathematically well-defined smoothness conditions over smaller patches of deformable surfaces and resorting to image-based collision detection tests, we developed an efficient collision detection method that achieves interactive rates while tracking self-interactions in highly deformable surfaces consisting of more that 50,000 elements. The method makes use of a novel technique for dynamically generating a hierarchy of cloth bounding boxes in order to perform object-level culling and image-based intersection tests using conventional graphics hardware support.

In modern urban areas, we often find a transportation network that follows a superimposed pattern. In this paper, we propose a novel method to generate a virtual traffic network based on (1) image-derived templates, and (2) a rule-based generating system. Using 2D images as input maps, various road maps with different patterns could be produced. This traffic network generating model adjusts itself intelligently in order to avoid restricted geographical areas or urban developments. The generative model follows closely directions of elevation and connects road ends in ways that allow various types of breakpoints.

We present a new subdivision scheme that is shown to improve the performance of ray tracing surfaces of revolution over Kajiya's (1983) classical work. This is based on a monotonic interval partitioning of a generatrix of a surface of revolution. The algorithm has a search complexity upper bound of O(log(m n)) for m monotonic intervals and n subdivisions for each interval and runs up to three times faster on large scenes. This method also suggests a novel hybrid bounding volume scheme that reduces this number of intersection tests between a ray and the actual object surface.

Geometric modeling and computational representations of shapes have been subject to intense research for more than three decades. Interestingly, these subjects are still at the heart of a continuous activity of research and development in computer graphics, virtual environments, image-based rendering, computer-aided geometric design and physical simulations. Currently, geometric and physically-based modeling still face two main challenges: (1) the identification of topological features, and (2) the representation of the modes of interaction between them, both in static and dynamic environments. Current methods have offered many different forms of associating abstract structures with analytical expressions. The variety of modeling tools, from combinational methods to analytic algebraic geometry, not only reflects the richness of ideas in this domain of study but also the desire to improve, enhance and simplify. It is within this realm that we introduce a new framework, called holorgraphic geometric modeling (HGM). This framework combines the advantages of the graph-theoretic representation of combinatorial structures with the analytical flexibility, expressional power and scalability of higher-order, multi-dimensional variables and operators in the form of holors. HGM not only complements the combinatorial structures in geometric modeling but also enhances and reveals new concepts and ideas in the process of developing robust, flexible and scalable domains of formulation for simplicial complexes, cellular spaces, and homotopy in general.

Accurate global visibility information has become essential in global illumination rendering. Combinatorial structures, such as the visibility complex, aspect graphs and the visibility skeleton, are the seeds of a new topic of research motivated by practical problems in computer graphics and global illumination rendering. This is now attracting interest from the computational geometry, combinatorics, CAD and robotics communities. The initial motivation has been to accurately represent shadow boundaries generated by occluding surfaces in the path of light from emitting surfaces (sources) to receiving surfaces (receivers). This article adds to the rich body of research in this field by presenting solutions to the problem of occluders with curved boundaries. This has profound implications in three main areas: (1) the generation, maintenance and search functions associated with the global visibility complex, (2) mesh generation, and (3) global illumination rendering. We present a new visibility complex based on the proposed visibility skeleton. The new visibility complex has been designed to handle not only polygonal elements but also surfaces with curved boundaries. We have found that, even in simple 3D scenes, the new visibility complex generates less polygons in the meshing phase and a more accurate global illumination

In interactive virtual environments and dynamic simulations, collisions between complex objects and articulated bodies may occur simultaneously at multiple points or regions of interference. Many solutions to the collision response problem are formulated based on the local pair‐wise contact dynamics. In this article, we present a new solution to the global interactions and dynamic response between multiple structures in a three‐dimensional environment. This is based on a new dynamic impulse graph that tracks the reaction forces through the entire system and gives a global view of all the interactions in a multibody system.

Object interactions are ubiquitous in interactive computer graphics, 3D object motion simulations, virtual reality and robotics applications. Most collision detection algorithms are based on geometrical object-space interference tests. Some algorithms have employed an image-space approach to the collision detection problem. In this paper we demonstrate an image-space collision detection process that allows substantial computational savings during the image-space interference test. This approach makes efficient use of the graphics rendering hardware for real-time complex object interactions. Copyright © 1999 John Wiley & Sons, Ltd.

Object interactions are ubiquitous in interactive computer graphics, 3D object motion simulations, virtual reality and robotics applications. Most collision detection algorithms are based on geometrical object space interference tests. Some algorithms have employed an image space approach to the collision detection problem. We demonstrate an image space collision detection process that allows substantial computational savings during the image space interference test. This approach makes efficient use of the graphics rendering hardware for real time complex object interactions.

In a virtual environment as well as in practice, collisions between complex surfaces and/or articulated bodies may occur simultaneously at multiple points of interference. Previous solutions to the collision response problem have been formulated based on pairwise rigid body dynamic interactions. These give only a local view of the dynamic response of each body. We present a new solution to the global interactions and dynamic response between multiple structures in a three-dimensional environment. This is based on a graph-theoretical framework that allows a systematic formulation of the equations of motion, gives a global view of all the interactions in a system, and suggests other approaches to the computation of dynamic response of a system.

The HolorGraphic modelling system (HGMS) is a general computational framework for the algebraic manipulation of complex mathematical models of both physical and non-physical systems. The lack of constraints on the operational algebra in HGMS provides both flexibility and expressive power for manipulating variables and functions. The underlying influence of the HolorGraphic framework is now being independently observed in seemingly unrelated fields of studies, from theoretical physics to the construction and analysis of curves and surfaces in computer aided geometric design. As it takes its own form above and beyond tensor theory, the HolorGraphic framework provides both a consistent mathematical basis of formulation as well as a visual understanding of the object of study. In this article, we illustrate the use of HGMS in the symbolic computation procedures using general holors and show the basic operations on these structures.

Although object manipulation is an important issue in VR, most applications fail to provide the natural interface of a dextrous hand that interacts with the objects in a physics-based realism. The paper presents a hybrid control approach that uses both kinematics and dynamics methods at different stages of picking to generate physically-accurate hand interactions in real time. This approach can simulate the natural picking interactions in various hand applications, including 3D mechanical design, assembly, surgical simulation, and telerobotics.

To provide fast display and manipulation of the 3D reconstruction of volumetric data, our system uses surface rendering which reduces 3D data sets to geometric primitives that have been implemented in the SGIs hardware, so a better performance of rendering speed is obtained. Our system supports both basic and advanced navigation features of medical data sets, including translating, rotating, zooming in/out, advanced cutting functions (transparent alignment, multi-plane cut, cut in any orientation and depth), classification of tissues, and centralization function. The 3D interface of our system provides users realistic view and feature-based manipulation of the medical data sets to extract useful information for medical training and education.

Article Free Access Share on Hardware-assisted virtual collisions Authors: George Baciu Computer Science Department, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong Computer Science Department, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong KongView Profile , Wingo Sai-Keung Wong View Profile , Hanqiu Sun View Profile Authors Info & Claims VRST '98: Proceedings of the ACM symposium on Virtual reality software and technologyNovember 1998 Pages 145–151https://doi.org/10.1145/293701.293720Published:02 November 1998Publication History 3citation300DownloadsMetricsTotal Citations3Total Downloads300Last 12 Months25Last 6 weeks5 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteeReaderPDF

Visualizing architectural designs with t,he aid of Virtual Reality (VR) technology is becoming a common task among architects, design engineers, contractors and customers due to a more realistic resemblance to the real look of the final constructions.In order to preview virtual buildings, one needs to reconstruct a VR model from the two-dimensional architectural drawings.In the conventional approach, a large amount of manual tasks are necessary while using some common 3D authoring tools avaliable in the market.This gives us the incentive to examine the process of modelling and reconstruction of 2D designs for use in practical applications.In this paper, we identify the main tasks in the conventional reconstruction process and give a semi-aut,omated solution that aids in the process of ext,rusion and modelling of large building complexes.

Collision detection between complex objects using rasterizing graphics hardware provides a rich ground of exploration for speeding up the interference detection algorithms and computing points of collision. We show that despite the limitations imposed by the current graphics rendering hardware, it is still possible to perform collision detection at rendering rates using only conventional graphics hardware without any enhancements.

The distinction between geometry and dynamic interactions is fundamental for the consistent dynamic analysis of physical systems. A unified treatment of such systems is possible when we adopt a hierarchical mathematical model with a consistent set of embedded abstractions. This new view is adopted in the general formulation strategy for obtaining a simplified dynamics model of mechanical systems. We show that there exists a consistent general extension from the model of constrained particle-mass systems (PMS) to the model of multibody systems (MBS) based entirely on graph-theoretic concepts.

With the aid of graph theory it is possible to combine the topological information of a constrained rigid-body system with the mathematical formulation of the equations of motion in a direct, systematic, and consistent procedural fashion that lends itself to a straightforward computer implementation. Such a graph-theoretical formulation for the dynamics analysis and simulation of three-dimensional constrained rigid-body systems is presented. It is shown that through the graph-theoretic approach, the topological information contributes directly to the systematic formulation procedure and also to an efficient, directly implementable computer algorithm. This algorithm can be implemented in a recursive fashion that reduces the computational cost substantially. The main components of the numerical procedure can be traced back directly to the system graphs of the formulation phase with very little effort. The resulting system of differential equations is left in its implicit form and solved numerically.>

The graph-theoretic models of a prismatic joint, a revolute joint, and an open kinematic chain are presented. The Denavit-Hartenberg representation of linkages is encompassed within the general framework of graph-theoretic system theory. The final mathematical model derived by this formalism is a system of differential and algebraic equations.

A set of mixed differential and algebraic equations (DAEs) which arises in the simulation of a robot manipulator is solved simultaneously using implicit integration. The dimension of the DAEs which have to be solved by LU factorization at each integration step can be reduced to the number of degrees of freedom by exploring the special structure of the Jacobian matrix of DAEs. The independent and dependent generalized coordinates are determined directly from the system topology. The simulation of a 6-R manipulator is given as an example.