PG2014short

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/7750

Browse

Now showing 1 - 20 of 26

Real-time Collision Detection with Two-level Spatial Hashing on GPU
(The Eurographics Association, 2014) Hong, Yang; Wu, Wen; Chen, Hui; John Keyser and Young J. Kim and Peter Wonka
In this paper, a two-level parallel spatial hashing method is presented for real-time collision detection of deformable objects based on modern GPU architecture. The second-level of spatial hashing is used to improve the culling efficiency. Moreover, a novel encoding method on GPU is proposed to compensate the inflexibility of the GPU memory system. It can efficiently determine the colliding pairs of primitives between deformable objects. The experimental results show that our method can perform high culling efficiency with low memory cost.
Geometry-Aware Image Completion via Multiple Examples
(The Eurographics Association, 2014) Wang, Guihang; Chen, Xuejin; Hu, Siyu; John Keyser and Young J. Kim and Peter Wonka
When browsing through photographs taken during a trip, it can be a distressing discovery to find many other bystanders captured within the frame. A visually compelling snapshot preserves the desired subject in the foreground, and eliminates irrelevant persons or objects. In this paper, we present a new image completion algorithm that retains the geometric consistency of the scene using Internet photo collections. Our key insight allows us to successfully fill the missing regions with content derived from related images with distinct features, or from visible parts of the input image with repetitive textures. The final composition for the missing regions is guided by a confidence map combining the two types of example images. We validate our algorithm on a variety of scenes which are challenging for state-of-the-art completion techniques. The completion results successfully preserve the geometry consistency for complex structures in a wide range of depths.
Perceptually Optimised Illumination for Seamless Composites
(The Eurographics Association, 2014) Chalmers, Andrew; Choi, Jong Jin; Rhee, Taehyun; John Keyser and Young J. Kim and Peter Wonka
Realistic illumination in composition is important for a seamless mixture between the virtual and real world objects in visual effects and mixed reality. The seamlessness is the measure of how perceivably apparent the synthetic object in the final composition is, and how indistinguishable it is from the photographed scene. Given that the ultimate receiver of image information is the human eye, the metric is determined by the Human Visual System (HVS). We conducted a series of psychophysical studies to observe and capture the thresholds of the HVS's ability to perceive illumination inconsistencies between the rendered subject and the photographed scene. Based on our observations, we find perceptually optimised thresholds for reducing resources across resolution and dynamic range of the radiance map (RM) for image based lighting (IBL). We evaluated our thresholds to illuminate virtual objects for seamless composition with photographed scenes.
Example-based Haze Removal with two-layer Gaussian Process Regressions
(The Eurographics Association, 2014) Fan, Xin; Gao, Renjie; Wang, Yi; John Keyser and Young J. Kim and Peter Wonka
Hazy images suffer from low visibility and contrast. Researchers have devoted great efforts to haze removal with the prior assumptions on observations in the past decade. However, these priors from observations can provide limited information for the restoration of high quality, and the assumptions are not always true for generic images in practice. On the other hand, visual data are increasing as the popularity of imaging devices. In this paper, we present a learning framework for haze removal based on two-layer Gaussian Process Regressions (GPR). By using training examples, the two-layer GPRs establish direct relationships from the input image to the depth-dependent transmission, and meanwhile learn local image priors to further improve the estimation. We also provide a method to collect training pairs for images of natural scenes. Both qualitative and quantitative comparisons on simulated and real-world hazy images demonstrate the effectiveness of the approach, especially when white or bright objects and heavy haze regions appear and existing dehazing methods may fail.
Scene Segmentation and Understanding for Context-Free Point Clouds
(The Eurographics Association, 2014) Spina, Sandro; Debattista, Kurt; Bugeja, Keith; Chalmers, Alan; John Keyser and Young J. Kim and Peter Wonka
The continuous development of new commodity hardware intended to capture the surface structure of objects is quickly making point cloud data ubiquitous. Scene understanding methods address the problem of determining the objects present in a point cloud which, dependant on sensor capabilities and object occlusions, is normally noisy and incomplete. In this paper, we propose a novel technique which enables automatic identification of semantically meaningful structures within point clouds acquired using different sensors on a variety of scenes. The representation model, namely the structure graph, with nodes representing planar surface segments, is computed over these point clouds to help with the identification task. In order to accommodate for more complex objects (e.g. chair, couch, cabinet, table), a training process is used to determine and concisely describe, within each object's structure graph, its important shape characteristics. Results on a variety of point clouds show how our method can quickly discern certain object types.
Incorporating Fiber Controls into FEM Model for Transversely Isotropic Materials
(The Eurographics Association, 2014) Jianping, Cai; Feng, Lin; Tsui, Lee Yong; Kemao, Qian; Soon, Seah Hock; John Keyser and Young J. Kim and Peter Wonka
Physically plausible deformable models based on continuum mechanics have been a hot topic in computer graphics for decades, and many models have been proposed to improve performance speed and stability. However, most of the existing models focus on isotropic materials, while elastic objects with complex anisotropic properties are less studied. Based on the observation that a large group of objects have specific internal structures (fibers) that determine their anisotropic behavior, we propose a fiber incorporated corotational FEM model that can approximate longitudinally anisotropic deformation. First, a fiber orientation field is used to establish local frames for each element; then, the orientation information is combined into the FEM model by adding local transformations on element stiffness matrices. This proposed model can provide a control for directable deformations, and yields realistic anisotropic deformations. Large deformations can be accommodated; meanwhile, with pre-computation it adds no computational cost to the existing corotational FEM model during simulation. Convincing experimental results and analytical comparisons are presented, together with an accompanying video demonstration.
Automatic Aesthetics-based Lighting Design with Global Illumination
(The Eurographics Association, 2014) Léon, Vincent; Gruson, Adrien; Cozot, Rémi; Bouatouch, Kadi; John Keyser and Young J. Kim and Peter Wonka
In computer graphics, lighting plays an important role in the appearance of a scene. A change in the configuration of light sources can lead to different aesthetics in the final rendered image. Lighting design becomes increasingly complex when using sophisticated global illumination techniques. In this paper, we present a new approach to automatically design the lighting configuration according to the aesthetic goal specified by the user as a set of target parameters. Target parameters are used to set up an objective function which is minimized using an optimization method. The results show that our method can be used to automatically design a lighting configuration that will give to the final image a classic photographic look.
Surface Mesh Segmentation and Reconstruction with Smooth Boundary Curves
(The Eurographics Association, 2014) Tsuchie, Shoichi; Higashi, Masatake; John Keyser and Young J. Kim and Peter Wonka
In mesh segmentation for industrial design objects, each segment defined by a region with closed boundary should correspond to its underlying surface constructed according to the designer's intention. In order to generate those segments from the scanned data, we propose a method in which (i) more suitable regions are extracted by the region splitting/merging processing with a new splitting scheme from the obtained clusters, (ii) smooth and consistent boundaries are generated as intersection or contacting curves between the adjacent underlying surfaces, and (iii) the region is reconstructed with high-quality triangle facets. We demonstrate the effectiveness of our method by applying it to the scanned data of the real-world industrial design objects.
Extraction and Transfer of Facial Expression Wrinkles for Facial Performance Enhancement
(The Eurographics Association, 2014) Shin, Il-Kyu; Öztireli, A. Cengiz; Kim, Hyeon-Joong; Beeler, Thabo; Gross, Markus; Choi, Soo-Mi; John Keyser and Young J. Kim and Peter Wonka
Creating realistic models and animations of human faces with adequate level of detail is very challenging due to the high complexity of facial deformations and the sensitivity of the human visual system to human faces. One way to overcome this challenge is using face scans captured from real facial performances. However, capturing the required level of detail is only possible with high-end systems in controlled environments. Hence, a large body of work is devoted to enhancing already captured or modeled low resolution faces with details extracted from high fidelity facial scans. Details due to expression wrinkles are of particular importance for realistic facial animations. In this paper, we propose a new method to extract and transfer expression wrinkles from a high resolution example face model to a target model for enhancing the realism of facial animations. We define multi-scale maps that capture the essential details due to expression wrinkles on given example face models, and use these maps to transfer the details to a target face. The maps are interpolated in time to generate enhanced facial animations. We carefully design the detail maps such that transfer and interpolation can be performed very efficiently using linear methods.
Integrating Occlusion Culling into LOD on GPU
(The Eurographics Association, 2014) Peng, Chao; John Keyser and Young J. Kim and Peter Wonka
Real-time rendering of complex 3D models is still a very challenging task. Recently, GPU-based level-of-detail (LOD) approaches have been proposed to fast decrease the complexity of a 3D model, but applying only LOD approaches is usually not sufficient to achieve highly interactive rendering rate for the complex model that contains hundreds of millions of triangles. Visibility culling, especially occlusion culling, needs to be introduced to further reduce the amount of triangles being rendered at each frame. In this paper, we present a novel rendering approach that seamlessly integrates occlusion culling with the LOD approach in a unified scheme towards the GPU architecture. The result shows that the integration significantly reduces the complexity of the 3D model and satisfies the demands of both memory efficiency and performance.
Finding Feature Similarities Between Geometric Trees
(The Eurographics Association, 2014) Mukherjee, Uddipan; Gopi, Meenakshisundaram; John Keyser and Young J. Kim and Peter Wonka
Geometric trees are graphs with no cycles in which the nodes have spatial co-ordinates and the edges are geometric curves. Many physical systems can be represented effectively using geometric trees, e.g. river beds, animal neurons, respiratory tracks of mammals etc. As these systems undergo structural metamorphosis, temporally or under the effect of some external stimulus, the underlying tree structures also change. Given two snapshots of structurally morphed trees, an algorithm for comparing them based on geometric and topological tree features is presented. Such comparison provides a wealth of information for interpreting the metamorphosis.
Visual Analysis of FPS Gameplay Data: From Game Design to Player Behavior
(The Eurographics Association, 2014) Li, Quan; Qu, Huamin; John Keyser and Young J. Kim and Peter Wonka
Gameplay data analysis has already become an important method for analyzing player behavior in games. Visualization is a promising way to explore and gain insight into the data. In this paper, we work closely with the game designers and user experience engineers to develop a visual analytic system to help them explore the gameplay data for a novel FPS (First-Person Shooter) game specific in the mainland China. We first come up with task specifications for such a system. After that, we propose a set of design goals for our system. VisFPS, is thus developed iteratively through a complete use-centered design process. The system is divided into two parts: Macro-View to deal with the overall gameplay data to discover patterns, and Micro-View to focus on a specific game match to recreate the game scene and use it to study player behavior and verify the game design intent.
Automatic 3D Posing from 2D Hand-Drawn Sketches
(The Eurographics Association, 2014) Gouvatsos, Alexandros; Xiao, Zhidong; Marsden, Neil; Zhang, Jian J.; John Keyser and Young J. Kim and Peter Wonka
Inferring the 3D pose of a character from a drawing is a non-trivial and under-constrained problem. Solving it may help automate various parts of an animation production pipeline such as pre-visualisation. In this paper, a novel way of inferring the 3D pose from a monocular 2D sketch is proposed. The proposed method does not make any external assumptions about the model, allowing it to be used on different types of characters. The 3D pose inference is formulated as an optimisation problem and a parallel variation of the Particle Swarm Optimisation algorithm called PARAC-LOAPSO is utilised for searching the minimum. Testing in isolation as well as part of a larger scene, the presented method is evaluated by posing a lamp and a horse character. The results show that this method is robust and is able to be extended to various types of models.
Quantitative Analysis of Voxel Raytracing Acceleration Structures
(The Eurographics Association, 2014) Chajdas, Matthäus G.; Westermann, Rüdiger; John Keyser and Young J. Kim and Peter Wonka
In this work, we provide a comprehensive analysis of GPU acceleration structures for voxel raytracing. We compare the commonly used octrees to BVH and KD trees, which are the standard in GPU triangle raytracing. Evaluating and analyzing of the behavior is complicated, as modern GPUs provide wide vector units with complex cache hierarchies. Even with sophisticated SIMD simulators, it is increasingly hard to model the hardware with sufficient detail to explain the observed performance. Therefore, instead of relying on SIMD simulation, we use hardware counters to directly measure key metrics like execution coherency on a modern GPU. We provide an extensive analysis comparing different acceleration structures for different raytracing scenarios like primary, diffuse and ambient occlusion rays. For different scenes we show that data structures commonly known for good performance, like KD-trees, are actually not suited for very wide SIMD units. In our work we show that BVH trees are most suitable for GPU raytracing and explain how the acceleration structure affects the execution coherency and ultimately performance, providing crucial information for the future design of GPU acceleration structures.
A Simple Artistic Rendering Method for Stereoscopic Images
(The Eurographics Association, 2014) Li, Dajin; Bai, Chengjie; John Keyser and Young J. Kim and Peter Wonka
General painterly rendering algorithms cannot be used to render stereo image pair separately because they may produce random noise and lead to inconsistency between binocular images. In this paper, a generic rendering framework of artistic stereoscopic images for virtual 3D scenes is presented. Firstly, an artistic image that contains entire visible surfaces of a 3D model from two eyes was rendered. During the projection transformation, the texture coordinates which correspond to vertices of the model were recorded. Then, the artistic image is mapped to object surface so as to obtain the stylized model. Finally, the stylized 3D model is projected to the left and right cameras for artistic stereoscopic images. The proposed rendering framework can guarantee the consistency between binocular images; moreover, it also can be applied to various existing artistic rendering algorithms.
Projecting Points onto Planar Parametric Curves by Local Biarc Approximation
(The Eurographics Association, 2014) Song, Hai-Chuan; Shi, Kan-Le; Yong, Jun-Hai; Zhang, Sen; John Keyser and Young J. Kim and Peter Wonka
This paper proposes a geometric iteration algorithm for computing point projection and inversion on surfaces based on local biarc approximation. The iteration begins with initial estimation of the projection of the prescribed test point. For each iteration, we construct a 3D biarc on the original surface to locally approximate the original surface starting from the current projection point. Then we compute the projection point for the next iteration, as well as the parameter corresponding to it, by projecting the test point onto this biarc. The iterative process terminates when the projection point satisfies the required precision. Examples demonstrate that our algorithm converges faster and is less dependent on the choice of the initial value compared to the traditional geometric iteration algorithms based on single-point approximation.
Saliency-driven Depth Compression for 3D Image Warping
(The Eurographics Association, 2014) Gu, Minjie; Hu, Shanfeng; Wang, Xiaochuan; Liang, Xiaohui; Shen, Xukun; Qin, Aihong; John Keyser and Young J. Kim and Peter Wonka
Current compression methods compress depth images by incorporating 2D features, which leads to a loss of the detail of the original 3D object in the recovered depth image. The main idea of this paper is to augment 2D features with 3D geometric information to preserve important regions of the depth image. Mesh saliency is used to represent the important regions of the 3D objects, and discontinuity edges are extracted to indicate the important regions of the depth image. We use mesh saliency to guide the adaptive random sampling to generate a random pixel sample of the depth image and then, combine this sample with the depth discontinuity edge to build the sparse depth representation. During the depth reconstruction, the depth image is recovered by using an up- and down-sampling schema with Gaussian bilateral filtering. The effectiveness of the proposed method is validated through 3D image warping applications. The visual and quantitative results show a significant improvement of the synthetic image quality compared with state-of-the-art depth compression methods.
Visibility Filtering for Producing Indirect Illumination
(The Eurographics Association, 2014) Chen, Yu-Jung; Yen, Chen-Yu; Chen, Yen-Yu; Chen, Wei-Chao; Chien, Shao-Yi; John Keyser and Young J. Kim and Peter Wonka
Indirect illumination is one of the most visually significant effects for a synthesized image. In order to render such effects at interactive rates, it is important to resolve the visibility between surfaces efficiently. In this work, we propose to approximate visibility by representing the scene through point proxies and hash them into voxels. The indirect illumination is then rendered in two stages. First, we cast sparse shadow rays to march and collect visible voxels for each low-resolution deferred pixel. The list of visible voxels are then processed using the proposed Visibility Filter (VF) to collect voxels missed from the first stage. Afterward, we properly weight the visible voxels as light sources to produce the full resolution image. Our method runs at interactive rates at 1.3 fps to 2.3 fps on GPUs with pleasing visual quality, in particular for glossy reflections.
PatchMove: Patch-based Fast Image Interpolation with Greedy Bidirectional Correspondence
(The Eurographics Association, 2014) Saito, Shunsuke; Sakamoto, Ryuuki; Morishima, Shigeo; John Keyser and Young J. Kim and Peter Wonka
In this paper, we present a method for the plausible interpolation of images. This method has several applications, such as for smooth view interpolation, low frame-rate video upsampling, and animation. The central idea is to quickly form dense correspondences using a patch-based nearest-neighbor search method called PatchMatch. However, the conventional PatchMatch method does not always find an accurate correspondence. This means that some patches do not find appropriate counterparts. Our method employs a greedy algorithm and an occlusion handling technique to correct inaccurate correspondences. Furthermore, our texture reconstruction method successfully reduces blurring effects. We demonstrate that our method significantly reduces the computation time required for interpolation, and show that the quality of reconstructed images is almost identical to that of those generated using state-of-the-art methods.
Automatic Garment Modeling From Front And Back Images
(The Eurographics Association, 2014) Huang, Lifeng; Gao, Chengying; John Keyser and Young J. Kim and Peter Wonka
We present a system which can automatically generate a realistic garment model from two images of an existing garment. Without the requirement of tailoring expertise and tedious operation, our method takes the front and back images of a real garment as input, and the system will make reasonable geometric modeling as well as physical simulation of the garment. Combining with mannequin's skeleton information, we propose a panel positioning method to place garment panels in appropriate positions. A key feature of our system is to automatically interpret sewn information, which effectively simplifies user interaction. In addition, panel deformation method based on mannequin's pose allows easy data capture. It extends the flexibility and utility of our method. The experiments demonstrate the effectiveness on generating models of various garment styles.

Browse

Browsing PG2014short by Issue Date

Results Per Page

Sort Options