Search

In this paper, we introduce a novel iterated random search method for large-scale texture synthesis and manipulations. Previous researches on texture synthesis and manipulation have reached a great achievement both on quality and performance. However, the cost of the popular exhaustive search-based methods is still high especially for large-scale and complex synthesis scenes. Our algorithm contributes great improvements on performances about 2---50 times over the typical patch-based synthesis methods. Texture patterns have been well-known formalized as a Markov Random Field (MRF) whose two hypotheses, stationarity and locality, drive our bold guess that a random sampling may just catch a good match and allows us to propagate the natural coherence in the neighborhood. Meanwhile, the iteration constantly updates the bad guesses to make our algorithm converge fast with the results in the state of the art. We also provide a simple theoretical analysis to compare our iterated randomized search model and the classical synthesis algorithms. Besides, this simple method turns out to work well in various applications as well, such as texture transfer, image completion and video synthesis.

W. Wu · E. Wu University of Macau, Macao, China Fig. 1 Comparison of different values of . From left to right, the first one is semi-Lagrangian advection while the others used of x , 3 x and 5 x for DBCMmethod. Smaller led to more detailed and turbulent surface. can be used for controlling the details. As level set was used for identifying the fluid domain, the underlying simulations could be slightly different with different values of . Even with high tolerance value, the result was still much better than that of semi-Lagrangian advection

In this paper, we present a new numerical method for advection in fluid simulation. The method is built on the Characteristic Mapping method. Advection is solved via grid mapping function. The mapping function is maintained with higher order accuracy BFECC method and dynamically reset to identity mapping whenever an error criterion is met. Dealing with mapping function in such a way results in a more accurate mapping function and more details can be captured easily with this mapping function. Our error criterion also allows one to control the level of details of fluid simulation by simply adjusting one parameter. Details of implementation of our method are discussed and we present several techniques for improving its efficiency. Both quantitative and visual experiments were performed to test our method. The results show that our method brings significant improvement in accuracy and is efficient in capturing fluid details.

In virtual environments, real-time simulation and rendering of dynamic fluids have always been the pursuit for virtual reality research. In this paper, we present a real-time framework for realistic fluid simulation and rendering on graphics processing unit. Because of the high demand for interactive fluids with larger particle set, the computational need is becoming higher. The proposed framework can effectively reduce the computational burden through avoiding the computation in inactive areas, where many particles with similar properties and low local pressure cluster together. While in active areas, the computation is fully carried out; thus, the fluid dynamics are largely preserved. Here, a robust particle classification technique is introduced to classify particles into either active or inactive. The test results have shown that the technique improves the time performance of fluid simulation largely. We then incorporate parallel surface reconstruction technique using marching cubes to extract the surfaces of the fluid. The introduced histogram pyramid-based marching cubes technique is fast and memory efficiency. As a result, we are able to produce plausible and interactive fluids with the proposed framework for large-scale virtual environments. Copyright © 2013 John Wiley & Sons, Ltd.

In this paper, we introduce a novel, real-time and robust hand tracking system, capable of tracking the articulated hand motion in full degrees of freedom (DOF) using a single depth camera. Unlike most previous systems, our system is able to initialize and recover from tracking loss automatically. This is achieved through an efficient two-stage k-nearest neighbor database searching method proposed in the paper. It is effective for searching from a pre-rendered database of small hand depth images, designed to provide good initial guesses for model based tracking. We also propose a robust objective function, and improve the Particle Swarm Optimization algorithm with a resampling based strategy in model based tracking. It provides continuous solutions in full DOF hand motion space more efficiently than previous methods. Our system runs at 40 fps on a GeForce GTX 580 GPU and experimental results show that the system outperforms the state-of-the-art model based hand tracking systems in terms of both speed and accuracy. The work result is of significance to various applications in the field of human---computer-interaction and virtual reality.

In the previous works, the real-time fluid-character animation could hardly be achieved because of the intensive processing demand on the character's movement and fluid simulation. This paper presents an effective approach to the real-time generation of the fluid flow driven by the motion of a character in full 3D space, based on smoothed-particle hydrodynamics method. The novel method of conducting and constraining the fluid particles by the geometric properties of the character motion trajectory is introduced. Furthermore, the optimized algorithms of particle searching and rendering are proposed, by taking advantage of the graphics processing unit parallelization. Consequently, both simulation and rendering of the 3D liquid effects with realistic character interactions can be implemented by our framework and performed in real-time on a conventional PC. Copyright © 2013 John Wiley & Sons, Ltd.

We present a method to make an accurate and dense reconstruction from the input of video captured by a free moving handheld camera in real time. By the method firstly, the positions of the camera and sparse 3D points are estimated by simultaneous localization mapping. Then the depth maps of selected reference frames are computed from corresponding camera bundles. Lastly a novel linear algorithm is also proposed to integrate all the depth maps into dense meshes partially. The main contributions of this paper are in the following points: the reference frames and corresponding camera bundles are able to be selected automatically, then accurate and smooth depth maps are generated in real time, and the depth maps are merged into a dense mesh by using a linear algorithm based on the error clouds optimization. Our algorithm is implemented on dual CPU and graphics processing unit in a parallel framework for improving the performance. Copyright © 2013 John Wiley & Sons, Ltd.

Because existing image abstraction systems can hardly incorporate with the changing light, we present an integrated image abstraction and relighting rendering system, which is based on a salient edge-guided luminance field optimization approach. For an input image, we first adopt a sparsity prior-based illumination decomposition method to remove its original illumination and have an intrinsic image. Meanwhile, we iteratively extract the salient edge inside by employing a message-passing strategy. Then, we simplify this image with a proposed salient edge-guided image abstraction optimization algorithm in the luminance field. Finally, we put forward a salient edge-guided image relighting optimization method to simulate the effect of dynamic lighting along different directions. Experiment results show that our system can artistically adjust the illumination of the abstracted image and makes it more vivid. Copyright © 2013 John Wiley & Sons, Ltd.

Cross-domain visual matching aims at finding visually similar images across a wide range of visual domains, and has shown a practical impact on a number of applications. Unfortunately, the state-of-the-art approach, which estimates the relative importance of the single feature dimensions still suffers from low matching accuracy and high time cost. To this end, this paper proposes a novel cross-domain visual matching framework leveraging multiple feature representations. To integrate the discriminative power of multiple features, we develop a data-driven, query specific feature fusion model, which estimates the relative importance of the individual feature dimensions as well as the weight vector among multiple features simultaneously. Moreover, to alleviate the computational burden of an exhaustive subimage search, we design a speedup scheme, which employs hyperplane hashing for rapidly collecting the hard-negatives. Extensive experiments carried out on various matching tasks demonstrate that the proposed approach outperforms the state-of-the-art in both accuracy and efficiency.

In this paper, we present a hybrid method, which integrates PIC/FLIP and vortex particle methods into a unified framework, to efficiently simulate vortex shedding that happens when fluids flow around internal obstacles. To improve efficiency and reduce the numerical dissipations, we first solve the governing equations on a coarse grid using PIC/FLIP, and then interpolate the intermediate results to a finer grid to obtain the base flow. When the regular particles in PIC/FLIP enter the boundary layer, if the specified conditions are satisfied to cause vortex shedding, they are selected as vortex particles by assigning additional vorticity related attributes. The vortex particle dynamics are controlled by the vorticity form of NS equations, and several efficient methods are proposed to solve them on the finer grid. Finally, the obtained turbulence flow is added to the base flow. As a result, we are able to simulate turbulent water with rich wake details around the internal obstacles.

This paper presents a novel GPU-based multiresolution rendering on sole-cube maps (SCMs), which is a variant of geometry images built upon spherical parameterization. Given spherical parametrization of a manifold mesh, the sphere domain is gnomonically projected to a closed cube, which constitutes the 6-chart sole-cube maps. A quadtree structure of SCMs and normal map atlas are then constructed by using the regular re-sampling. Then, by packing the quadtree nodes into the SCMs texture atlas, a new parallel multiresolution rendering is processed on the latest GPU in two rendering passes: the multiresolution node selection in fragment shader; the triangulation in vertex shader followed by the node culling operation in geometry shader. The proposed approach generates adaptive mesh surfaces dynamically, and can be fully implemented in GPU parallelization. The proposed scheme alleviates the computing load of multiresolution mesh refinement on CPU, and our GPU-based multiresolution rendering is demonstrated with a variety of examples. Our user study confirmed that the visual quality of the SCMs multiresolution rendering, in comparison with the meshes/geometry images rendering, is also highly efficient especially for complex models in large-scale virtual environment.

This paper describes a new video stylization approach that achieves non-photorealistic rendering effects by using highly efficient spatial-temporal Gabor filtering. An edge extraction algorithm is developed to detect long coherent edges, to which the human visual system is sensitive. A nonlinear diffusion is then applied to remove unimportant details. Our approach extends the optical flow computation for constructing the Gabor flow to represent pixel similarity, and to preserve the temporal coherence when applied to video sequences. In particular, our video stylization is designed in a spatiotemporal manner to achieve temporal coherence in resulting animations. Real-time performance is achieved through the highly parallel implementation on modern graphics hardware (GPU). Therefore, our video stylization can be naturally applied to real-time video communication and interactive video-based rendering. The experimental results have demonstrated the high-quality production of our real-time video stylization.

Boundary priors have been extensively studied in salient region detection problems over the past few decades. Although several models based on the boundary prior have achieved good detection performance, there still exist drawbacks. The most common one is that they fail to detect the salient object when the background is complex or the salient object touches the image boundary. In this paper, we propose a novel model to detect the salient region. It is based on background cues and one complementary cue, that is, a foreground cue. A saliency score is obtained via solving an energy optimization problem which takes both the background cue and foreground cue into consideration. Extensive experiments, including both quantitative and qualitative evaluations on five widely used datasets, demonstrate the superiority of our proposed model to several other state-of-the-art models.

This paper presents a novel approach to fast rendering realistic meta-balls in view space. Meta-ball rendering is widely used in computer graphics to represent liquid drops, and especially important to particle-based fluid dynamics for the rendering on the final stage. Traditionally, meta-balls are procedurally modeled in object space according to implicit surface extraction and rendered as meshes. Instead of such consuming geometry-based methods, our algorithm is executed at screen-space level, and it utilizes the features of GPU, which demands low overheads and thus is much faster. Compared with other screen-space methods of particle rendering, our method obeys the basic principle of implicit surface. Consequently, real-time meta-ball rendering can be achieved very efficiently in a significantly high quality.

Although the stereo matching problem has been extensively studied during the past decades, automatically computing a dense 3D reconstruction from several multiple views is still a difficult task owing to the problems of textureless regions, outliers, detail loss, and various other factors. In this paper, these difficult problems are handled effectively by a robust model that outputs an accurate and dense reconstruction as the final result from an input of multiple images captured by a normal camera. First, the positions of the camera and sparse 3D points are estimated by a structure-from-motion algorithm and we compute the range map with a confidence estimation for each image in our approach. Then all the range maps are integrated into a fine point cloud data set. In the final step we use a Poisson reconstruction algorithm to finish the reconstruction. The major contributions of the work lie in the following points: effective range-computation and confidence-estimation methods are proposed to handle the problems of textureless regions, outliers and detail loss. Then, the range maps are merged into the point cloud data in terms of a confidence-estimation. Finally, Poisson reconstruction algorithm completes the dense mesh. In addition, texture mapping is also implemented as a post-processing work for obtaining good visual effects. Experimental results are presented to demonstrate the effectiveness of the proposed approach.

Video decolorization has wide application, such as in film industry. However, previous works on video decolorization focus on simply reusing the image decolorization method, which is time-consuming and may lose temporal coherence. This paper proposes a novel video decolorization method --- defining decolorization proximity, regarded as a criterion to determine the specific method, to do the decolorization and further preserving temporal coherence well. Experimented with vast diverse color videos, the experiment result shows that our framework can provide high efficiency and effectively avoid visual flicker and incoherence.

Cloud simulation is important for creating images of outdoor scenes. However, the complexity of this natural phenomenon makes the simulation of large-scale clouds difficult in real time. In this paper, we present a new method for 3D cloud simulation in which cloud animation is simplified and simulated by approximating Lennard-Jones Potential. To solve the N-body problem in Lennard-Jones Potential, we minimized the interaction between particles by dividing the simulation space into many cells and we defined a cutoff distance to perform calculation between neighboring particles. Additionally, a separate distance is introduced between particles to maintain the stability in the Lennard-Jones system. Our experimental results demonstrate that our method is computationally inexpensive and suitable for real time applications where large-scale simulation of clouds is required.

In this paper, we propose a novel method for the metamorphosis between two different images. By the approach, the transition sequence is generated by stitching two forward and backward warped sequences in a three-dimensional space along transition surface. In contrast to the traditional methods by blending two warped images at each intermediate frame, we continuously warp images on opposite direction without blending until the two warped images match in a three-dimensional space leading to a better transition in quality. Furthermore, for each pixel, we make decision of choosing a given input image best suitable so as to produce plausible in-between images to prevent from ghost and blur. By our scheme, the transition surface is computed by minimizing an energy function in terms of graph-cut optimization. Depending on the transition surface, a warp function is proposed to create a smooth and clear transformation. We demonstrate the advantage of our framework by performing transformation test to various kinds of image couples.

Line drawing plays an important role in many image-based non-photorealistic applications. However, most existing approaches use a grayscale edge detector for line extraction, so that only luminance differences between nearby image pixels is taken into account, but the chrominance differences is ignored. This leads to the undesirable consequence that visually significant edges in adjacent regions with different colors of similar luminance cannot be detected. To address this limitation, we present a novel enhanced line drawing method based on a flow-based difference-of-Gaussians (FDoG) filter. Because of an inherent property of the thresholded DoG edge model, captured lines may appear dislodged from the true edges in the image. To this end, we provide a gradient-guided warping technique so that smooth and coherent lines can be extracted in the correct location. The GPU implementation of the proposed algorithms allows real-time performance, and experimental examples with various color images demonstrate the method’s superior qualitative performance over previous approaches.

This paper presents a non-photorealistic rendering technique for stylizing a photograph in the line drawing style. We first construct a smooth and direction-enhancing edge flow field from the eigenvectors of the smoothed structure tensor, and then we use the flow field to guide the line drawing process. In particular, we develop a new operator for detecting step edges, which outperforms the existing edge detectors in terms of feature preservation and edge localization. Our approach works by applying the proposed detector in the direction perpendicular to the edge flow tangent and then smoothing the intermediate results along the edge flow curve. Optionally, an anisotropic nonlinear filter with an elliptical kernel is incorporated into the algorithms to extract the line edges, which may extend our technique further for creating an image to convey a hand-painting style. The presented algorithms are all highly parallel, allowing a real-time performance with GPU implementation. Experimental results show that our approach can produce more attractive and impressive line illustrations with a variety of photographs.

This paper presents a finite element method-based framework for an object withering simulation modeled with heterogeneous material, such as fruits drying or decay. We introduce diffusion procedures for both the moisture content and decay spread, which are solved directly on a tetrahedral mesh representation of the fruit flesh. Then, we use the moisture content to control shrinkage through the initial strain, which is integrated into the Lagrangian dynamic equation, and solved with the finite element method. For the complex structure of the object, another fine triangle mesh is used to represent the skin, and its deformation is solved by a thin shell technique. To couple the motion between different layers of the fruit, a tracking force is used to pull the skin and drive its deformation together with the volume mesh. In comparison with the previous work, our method provides temporally and spatially varying parameters to model the complex phenomena of object withering. Moreover, the water diffusivity can also be given by user input to present various material properties of the cut section and skin-covered area. Our algorithm is easy to implement and highly efficient in generating a realistic appearance for the withering effect. For a medium-scale model, we can achieve interactive simulation. Copyright © 2012 John Wiley & Sons, Ltd. (On the basis of a finite element method framework, the dynamic withering procedure is computed with diffusion of moisture content and decay spread. The material properties are controlled by moisture content, and the shrinkage also depends on moisture content through the initial strain. The heterogeneous deformation is simulated realistically and efficiently with our approach.)

This paper presents a novel approach for simulating the dynamic coupling between a tree and raindrops based on physical deformation and fluid simulation. By the approach, tree animation in the rain can be simulated in a two-resolution way: branch motion and leaf motion. The branch is represented by the Euler–Bernoulli beam model, and the leaf petiole is represented by the three-prism elastic model. Interaction coupling liquid motion on the hydrophilic surface with a flexible petiole is well implemented by a special design. To simplify the computation process, instead of the computation-intensive three-dimensional Navier–Stokes equations, shallow water equations are used to simulate the water dynamics together with the whole leaf deformation. Simulation has been also made to various phenomena incurred from the interactive coupling. These include, among others, part of impacting raindrops splashing into the air with the remaining flowing along the slant of the leaf and merging into larger ones or hanging on the blade boundary, with the leaf rebounding and vibrating after the drops fall off the leaf. A level-of-detail approach is exploited to accelerate rendering in views of different distances. The experimental results illustrate that the approach can be applied to efficiently generate realistic details of the interactive coupling between a tree and raindrops. Copyright © 2012 John Wiley & Sons, Ltd.

We present a new particle-based method for granular flow simulation. In the method, a new elastic stress term, which is derived from a modified form of the Hooke's law, is included in the momentum governing equation to handle the friction of granular materials. Viscosity force is also added to simulate the dynamic friction for the purpose of smoothing the velocity field and further maintaining the simulation stability. Benefiting from the Lagrangian nature of the SPH method, large flow deformation can be well handled easily and naturally. In addition, a signed distance field is also employed to enforce the solid boundary condition. The experimental results show that the proposed method is effective and efficient for handling the flow of granular materials, and different kinds of granular behaviors can be well simulated by adjusting just one parameter.

This paper presents a traffic-oriented modeling and simulation framework. With this tool, users can create a traffic scene quickly and easily, and then drive the vehicle in the created scene with physics simulated. Especially, the vehicle could be driven with vision technology, which is equipped with a webcam and a paper printed with controlling markers. And the paper is used as the virtual wheel to control the motion of the vehicle. At such low cost, many people can enjoy the driving fun; it also provides an easy and interesting way for driving video games with promising business values.

Beam tracing combines the flexibility of ray tracing and the speed of polygon rasterization. However, beam tracing so far only handles linear transformations; thus, it is only applicable to linear effects such as planar mirror reflections but not to non-linear effects such as curved mirror reflection, refraction, caustics and shadows. In this paper, we introduce non-linear beam tracing to render these non-linear effects. Non-linear beam tracing is highly challenging because commodity graphics hardware supports only linear vertex transformation and triangle rasterization. We overcome this difficulty by designing a non-linear graphics pipeline and implementing it on top of a commodity GPU. This allows beams to be non-linear where rays within the same beam do not have to be parallel or intersect at a single point. Using these non-linear beams, real-time GPU applications can render secondary rays via polygon streaming similar to how they render primary rays. A major strength of this methodology is that it naturally supports fully dynamic scenes without the need to pre-store a scene database. Utilizing our approach, non-linear ray tracing effects can be rendered in real-time on a commodity GPU under a unified framework.

This paper presents a new stylized augmented reality (AR) framework which can generate line drawing and abstracted shading styles. In comparison with the state-of-art work, our framework can significantly improve both the visual immersion of a single frame and the temporal coherence of augmented video streams in real time. In our framework, we first render virtual objects over the input camera images and then uniformly process the combined contents with stylization techniques. For generating line drawing stylization, we first propose a specially designed shading method to render the virtual objects, and then use an adapted Flow-based anisotropic Difference-of-Gaussion (FDoG) filter to yield the high-quality line drawing effect. For generating the abstracted stylization, a focus-guided diffusion filter and a soft color quantization operator are sequentially applied to the augmented image, and then the processed result is combined with the detected edges to produce the final abstraction effect. The presented algorithms are all sympathetic to highly parallel processing, allowing a real-time performance on contemporary graphics hardware.

The recently proposed matting Laplacian (Levin et al., IEEE Trans. Pattern Anal. Mach. Intell. 30(2):228–242, 2008) has been proven to be a state-of-the-art method for solving the image matting problem. Using this method, matting is formulated as solving a high-order linear system which is hard-constrained by the input trimap. The main drawback of this method, however, is the high computational cost. As the size of the input image increases, the matting Laplacian becomes expensive to solve in terms of both memory and computational time. In this paper we propose a GPU-based matting Laplacian solution which is dramatically faster than a conventional CPU solution, and at the same time largely reduces the memory consumption, making this method practical for the first time for high resolution image matting. To achieve this end, we employ a novel hierarchical windowing scheme to approximate the global optimal solution by solving a serial of local regions at multiple scales. We further employ a GPU-based local solver which can efficiently evaluate local solutions under various boundary conditions. Experimental results show that our system in general is more than two orders of magnitude faster than traditional CPU-based solvers, with about 80% less memory footprint.

We present a new model, called the dual-microfacet, for those materials such as paper and plastic formed by a thin, transparent slab lying between two surfaces of spatially varying roughness. Light transmission through the slab is represented by a microfacet-based BTDF which tabulates the microfacet's normal distribution (NDF) as a function of surface location. Though the material is bounded by two surfaces of different roughness, we approximate light transmission through it by a virtual slab determined by a single spatially-varying NDF. This enables efficient capturing of spatially variant transparent slices. We describe a device for measuring this model over a flat sample by shining light from a CRT behind it and capturing a sequence of images from a single view. Our method captures both angular and spatial variation in the BTDF and provides a good match to measured materials.

In this paper, we propose a novel conception of differential geometry images (DGIM), encapsulating differential coordinates to traditional geometry images. DGIM preserves the local geometric characteristics in many graphics applications, such as model remeshing and morphing. The traditional geometry images using Cartesian coordinates require normal maps for correctly rendering models, because they neglect the existence of local geometric details in the image structure, which leads us to compute the normals and curvatures imprecisely. Using our differential geometry images, normals can be reconstructed easily and correctly thereafter normal maps are no longer required. In addition, DGIM can be easily applied to mesh morphing due to its regular topology and well-preserved local details. In this paper, we also demonstrate a variety of plausible mesh morphing results based on DGIM in shape space.

In this paper, we present a hybrid representation of image-based models combining the textured planes and the hierarchical points. Taking a set of depth images as input, our method starts from classifying input pixels into two categories, indicating the planar and non-planar surfaces respectively. For the planar surfaces, the geometric coefficients are reconstructed to form the uniformly sampled textures. For nearly planar surfaces, some textured planes, called lumiproxies, are constructed to represent the equivalent visual appearance. The Hough transform is used to find the positions of these textured planes, and optic flow measures are used to determine their textures. For remaining pixels corresponding to the non-planar geometries, the point primitive is applied, reorganized as the OBB-tree structure. Then, texture mapping and point splatting are employed together to render the novel views, with the hardware acceleration.

This paper presents an interactive graphics processing unit (GPU)-based relighting system in which local lighting condition, surface materials and viewing direction can all be changed on the fly. To support these changes, we simulate the lighting transportation process at run time, which is normally impractical for interactive use due to its huge computational burden. We greatly alleviate this burden by a hierarchical structure named a transportation tree that clusters similar emitting samples together within a perceptually acceptable error bound. Furthermore, by exploiting the coherence in time as well as in space, we incrementally adjust the clusters rather than computing them from scratch in each frame. With a pre-computed visibility map, we are able to efficiently estimate the indirect illumination in parallel on graphics hardware, by simply summing up the radiance shoots from cluster representatives, plus a small number of operations of merging and splitting on clusters. With relighting based on the time-varying clusters, interactive update of global illumination effects with multi-bounced indirect lighting is demonstrated in applications to material animation and scene decoration.

A new patch-based texture synthesis method is presented in this paper. By the method, a set of patches that can be matched with a sampled patch for growing textures effectively, called the matching compatibility between patches, is generated first for each patch, and the set is further optimized by culling the patches that may cause synthesis conflicts. In this way, similarity measurement calculation for selecting suitable patches in texture synthesis can be greatly saved, and synthesis conflicts between neighbouring patches are substantially reduced. Furthermore, retrace computation is integrated in the synthesis process to improve the texture quality. As a result, the new method can produce high quality textures as texture optimization, the best method to date for producing good textures, and run in a time complexity linear to the size of the output texture. Experimental results show that the new method can interactively generate a large texture in 1024 × 1024 pixels, which is very difficult to achieve by existing methods.

Bao, K.; Wu, E. H. Chinese Acad Sci, Inst Software, State Key Lab Comp Sci, Beijing, Peoples R China. Fang, H. S.; Wei, J. A.; Zhang, H.; Zheng, L. L. SUNY Stony Brook, Dept Mech Engn, Stony Brook, NY 11794 USA. Zheng, L. L. Tsinghua Univ, Sch Aerosp, Beijing 100084, Peoples R China. Bao, K. Chinese Acad Sci, Grad Univ, Beijing, Peoples R China. Zhang, H. Tsinghua Univ, Dept Engn Phys, Ctr Publ Safety Res, Beijing 100084, Peoples R China. Wu, E. H. Univ Macau, Fac Sci & Technol, Macao, Peoples R China.

Due to high computational expense of simulating the physical model of atmospheric effect, rendering realistic natural light phenomena in real time is a challenging work. In this paper, we present a new approach for dynamic light shaft construction and rendering. Unified volumes construction is proposed for both dynamic light shaft and shadow, which is embedded into an efficient method of rendering natural light shafts with atmospheric scattering. We present the analytic formula of scattering computation without numerical integration and its implementation on current GPU. Our approach can simulate not only the lighting effect with single light source but also multiple parallel light sources according to the physical model of skylight and sunlight.

This system represents furry surfaces as nonuniform layers of texture slices, automatically adjusting the layers to achieve efficient, high-quality rendering. It employs layered shadow maps to simulate self-shadowing. Interactive tools let users intuitively create and edit furry objects and instantly view the rendered objects.

The State Key Laboratory of Computer Science (SKLCS) is committed to basic research in computer science and software engineering. The research topics of the laboratory include: concurrency theory, theory and algorithms for real-time systems, formal specifications based on context-free grammars, semantics of programming languages, model checking, automated reasoning, logic programming, software testing, software process improvement, middleware technology, parallel algorithms and parallel software, computer graphics and human-computer interaction. This paper describes these topics in some detail and summarizes some results obtained in recent years.

The paper presents a new algorithm for point-in-polygon queries. By decomposing a polygon into a set of convex polygons and then constructing a BSP tree, the algorithm performs an inclusion query against the polygon in two steps. In the first step, it finds the convex polygon that is most likely to contain the query point through the BSP tree, and then in the second step, it tests whether the query point is in the convex polygon by employing advanced techniques for point-in-convex-polygon queries. The new algorithm is immune of singularity and has a performance of conducting a query by time demand in O(logn), and storage requirement in O(n). The time complexity of the preprocessing is O(nlogn) for convex decomposition and BSP tree construction. The new algorithm has its performance in various aspects comparable to the state-of-art algorithms based on trapezoidation. Furthermore, as the number of convex polygons is always much fewer than the number of trapezoids in partitioning a polygon, and a high-quality search tree can always be constructed to manage the convex polygons, the new algorithm can run much faster than the trapezoidation-based algorithms, up to over double folds in maximum.

In the fluid simulation, the fluids and their surroundings may greatly change properties such as shape and temperature simultaneously, and different surroundings would characterize different interactions, which would change the shape and motion of the fluids in different ways. On the other hand, interactions among fluid mixtures of different kinds would generate more comprehensive behavior. To investigate the interaction behavior in physically based simulation of fluids, it is of importance to build physically correct models to represent the varying interactions between fluids and the environments, as well as interactions among the mixtures. In this paper, we will make a simple review of the interactions, and focus on those most interesting to us, and model them with various physical solutions. In particular, more detail will be given on the simulation of miscible and immiscible binary mixtures. In some of the methods, it is advantageous to be taken with the graphics processing unit (GPU) to achieve real-time computation for middle-scale simulation.

The progressive mesh representation is in general introduced for a continuous level of detail, and with it smooth transition between meshes can be achieved. However, the method is impractical for realtime simplification. Despite that, the realtime generation of progressive meshes is often required for applications such as in a robot simulation system or a flight simulation system where models are constantly modified by the actions of object collision, object destruction and reconstruction, etc. A new method is presented in this paper to tackle realtime simplification through progressive meshes. By the algorithm only the current mesh and readily accessible local information of each vertex, such as vertex position, edge length, neighboring vertices, adjacent faces, and face normals, are considered at each simplification step. All edge-collapse costs are calculated and sorted into a binary tree using the heap sort algorithm at the initial stage. At the iterative stages only the neighboring affected costs need be recalculated and rearranged in the binary tree. Other properties, such as vertex color and face texture, are processed in the same way as the geometry. The algorithm greatly improves the time performance under the constraint that the quality of the generated meshes be acceptable. Tests show that the algorithm is viable in a realtime simplification for medium-scale virtual models on PC platforms.

We introduce an adaptive hybrid level set surface tracking for fluid simulation in this paper. To improve the accuracy of the convectional level set method, we propose a method that combines the higher accuracy interpolation and octree grid to improve the accuracy while still maintain good efficiency. Adaptive gradient-augmented level set is adopted for higher accuracy interpolation and two strategies are proposed to guide the building of the octree structure. We demonstrate our method with several level set tests, and couple it to several existing fluid simulators. The results show that our method facilitates significant improvements in accuracy and efficiency, and is effective for free surface fluid simulation in capturing fluid details.

In this paper, we present an approach of a new surface tension formulation for Smoothed particle hydrodynamics (SPH). Our method of the surface tension formulation is decomposed into three main processes: (1) mesh smoothing, (2) surface tension computation and (3) surface tension transfer. Firstly, we exploit a Lagrangian operator to smooth an initial three-dimensional discrete interfacial surface mesh generated from the fluid particles using the Marching Cubes method, and then the surface mesh should be scaled in a volume-preserved way and its center should make a translation motion to its original position, thus we gain a new volume-preserved smoothed mesh which is abbreviated as smoothed mesh. Secondly, surface tensions strengths on the vertices of the interfacial surface mesh are computed according to the offsets from the original surface mesh to the smoothed mesh. Finally, we transfer the surface tension strengths above from the mesh vertices onto their neighbor particles of SPH in a conservation way. The proposed approach of surface tension solver is simple and straightforward to be plugged into a standard SPH solver. Experimental results show that it is effective and efficient to produce realistic fluid simulations, especially the phenomena associated with strong surface tension.

Accurate video object segmentation is still a difficult task so far due to the following reasons: the inseparable color between foreground and background (F/B) scenes is hard to segment only by color model, Occlusion/disocclusion caused by large movement, new exposed regions and topology change often breaks the continuity of video, Lastly, accurately building color model plays a critical role in obtaining a satisfactory result, which has often been ignored by people. In our method, all these problems have been considered: A kind of motion prediction method based on local coherence is adopted to separate the inseparable color. Then a self-adapting distance support model is used to build the object color by which our model can become much more robust against occlusion/disocclusion and color confusion. At last the final probability is generated by fusing all the clues, and the binary segmentation is finished by 3D Graph-Cut optimization. The experimental results are presented to demonstrate the effectiveness of the proposed method at achieving high quality results, as well as the robustness of the proposed method against several challenging test inputs.

Multi-task learning has been proposed to improve the generalization performance by learning multiple tasks jointly. One challenge for this learning paradigm is to effectively seek the shared information across multiple tasks. In this paper, we propose a novel multi-task learning method to adaptively share information. Unlike many existing multi-task learning methods which impose strong assumptions on task related-ness, our method captures the relationships among tasks and identifies the disparities of each task simultaneously, thus can flexibly exploit the shared information. Moreover, we apply it to fine-grained categorization problem, which usually suffers from the difficulties of insufficient training data and high inter-class similarity. The experimental results on two widely used datasets show the superiority of our method compared with some state-of-the-art methods.

The simulation of fluid mixing under the Eulerian framework often suffers from numerical dissipation issues. In this paper, we present a mass-preserving convection scheme that offers direct control on the shape of the interface. The key component of this scheme is a sharpening term built upon the diffusive flux of a user-specified kernel function. To determine the thickness of the ideal interface during fluid mixing, we perform theoretical analysis on a one-dimensional diffusive model using the Fick's law of diffusion. By explicitly controlling the interface thickness using a spatio-temporally varying kernel variable, we can use our scheme to produce realistic fluid mixing effects without numerical dissipation artifacts. We can also use the scheme to control interface changes between two fluids, due to temperature, pressure, or external energy input. This convection scheme is compatible with many advection methods and it has a small computational overhead.