Depth camera such as Microsoft Kinect, is much cheaper than conventional 3D scanning devices, and thus it can be acquired for everyday users easily. However, the depth data captured by Kinect over a certain distance of extreme low quality. In this paper, we present novel system capturing full human body models using multiple Kinects. To avoid interference phenomena, use two Kinects to capture upper part lower respectively without overlapping region. A third used middle from opposite...

Abstract We present a novel approach to parameterize mesh with disk topology the plane in shape‐preserving manner. Our key contribution is local/global algorithm, which combines local mapping of each 3D triangle plane, using transformations taken from restricted set, global “stitch” operation all triangles, involving sparse linear system. The can be variety families, e.g. similarities or rotations, generating different types parameterizations. In first case, parameterization tries force 2D...

3D printers have become popular in recent years and enable fabrication of custom objects for home users. However, the cost material used printing remains high. In this paper, we present an automatic solution to design a skin-frame structure purpose reducing given object. The frame is designed by optimization scheme which significantly reduces volume guaranteed be physically stable, geometrically approximate, printable. Furthermore, number struts minimized solving l 0 sparsity optimization....

We propose a novel data-driven technique for automatically and efficiently generating floor plans residential buildings with given boundaries. Central to this method is two-stage approach that imitates the human design process by locating rooms first then walls while adapting input building boundary. Based on observations of presence living room in almost all plans, our designed learning network begins positioning continues iteratively other rooms. Then, are determined an encoder-decoder...

We introduce FPConv, a novel surface-style convolution operator designed for 3D point cloud analysis. Unlike previous methods, FPConv doesn't require transforming to intermediate representation like grid or graph and directly works on surface geometry of cloud. To be more specific, each point, performs local flattening by automatically learning weight map softly project surrounding points onto 2D grid. Regular can thus applied efficient feature learning. easily integrated into various...

In this paper, we propose HeadNeRF, a novel NeRF-based parametric head model that integrates the neural radiance field to representation of human head. It can render high fidelity images in real-time on modern GPUs, and supports directly controlling generated images' rendering pose various semantic attributes. Different from existing related models, use fields as 3D proxy instead traditional textured mesh, which makes HeadNeRF is able generate images. However, computationally expensive...

Abstract Aesthetic images evoke an emotional response that transcends mere visual appreciation. In this work we develop a novel computational means for evaluating the composition aesthetics of given image based on measuring several well‐grounded guidelines. A compound operator crop‐and‐retarget is employed to change relative position salient regions in and thus modify image. We propose optimization method automatically producing maximally‐aesthetic version input validate performance show its...

We present an easy-to-use image retouching technique for realistic reshaping of human bodies in a single image. A model-based approach is taken by integrating 3D whole-body morphable model into the process to achieve globally consistent editing effects. novel body-aware warping introduced reliably transfer effects from image, even under moderate fitting errors. Thanks parametric nature model, our parameterizes degree small set semantic attributes, such as weight and height. It allows easy...

The joint bilateral filter is a variant of the standard filter, where range kernel evaluated using guidance signal instead original signal. It has been successfully applied to various image processing problems, it provides more flexibility than achieve high quality results. On other hand, its success heavily dependent on signal, which should ideally provide robust estimation about features output Such not always easy construct. In this paper, we propose novel mesh normal filtering framework...

Abstract We present a novel algorithm for automatically co‐segmenting set of shapes from common family into consistent parts. Starting over‐segmentations shapes, our approach generates the segmentations by grouping primitive patches directly and obtains their correspondences simultaneously. The core is to compute an affinity matrix where each entry encodes similarity between two patches, which measured based on geometric features patches. Instead concatenating different one feature...

3D face reconstruction from a single image is classical and challenging problem with wide applications in many areas. Inspired by recent works animation RGB-D or monocular video inputs, we develop novel method for reconstructing faces unconstrained 2D images using coarse-to-fine optimization strategy. First, smooth coarse generated an example-based bilinear model aligning the projection of landmarks detected input image. Afterward, local corrective deformation fields, refined photometric...

We present a complete system to semantically decompose and reconstruct 3D models from point clouds. Different than previous urban modeling approaches, our is designed for residential scenes, which consist of mainly low-rise buildings that do not exhibit the regularity repetitiveness as high-rise in downtown areas. Our first automatically labels input into distinctive categories using supervised learning techniques. Based on semantic labels, objects different are reconstructed with...

The fifth generation mobile networks will be developed to improve area spectral and energy efficiency, provide uniform user experience. Hyper-dense small cell deployment can move devices closer the wireless network satisfy 5G system requirements. main challenge of this results from random deployment, dynamic on-off, flexible connection cellular core networks, flat architecture systems. Therefore, conventional planning radio resource management, which depend on a central control node, cannot...

Abstract We propose a fast method for 3D shape segmentation and labeling via Extreme Learning Machine (ELM). Given set of example shapes with labeled segmentation, we train an ELM classifier use it to produce initial test shapes. Based on the compute final smooth through graph‐cut optimization constrained by super‐face boundaries obtained over‐segmentation active contours computed from segmentation. Experimental results show that our achieves comparable against state‐of‐the‐arts, but reduces...

Many computer graphics problems require computing geometric shapes subject to certain constraints. This often results in non-linear and non-convex optimization with globally coupled variables, which pose great challenge for interactive applications. Local-global solvers developed recent years can quickly compute an approximate solution such problems, making them attractive choice applications that prioritize efficiency over accuracy. However, these suffer from lower convergence rate, may...

In this paper, we propose StereoPIFu, which integrates the geometric constraints of stereo vision with implicit function representation PIFu, to recover 3D shape clothed human from a pair low-cost rectified images. First, introduce effective voxel-aligned features vision-based network enable depth-aware reconstruction. Moreover, novel relative z-offset is employed associate predicted high-fidelity depth and occupancy inference, helps restore fine-level surface de-tails. Second, structure...

On the basis of bottom-up spillover theory, this study explores influence wellness tourism experience on tourists’ well-being. Considering tourists as research participants, tourist satisfaction, gender, and age were selected mediating moderating variables, respectively, to mechanism between A questionnaire survey was sent 445 participants who engaged in China, an empirical analysis conducted using structural equation model. The results showed that entertainment, esthetic, escape can...

While many 3D objects exhibit various forms of global symmetries, prominent intrinsic symmetries which exist only on parts an object are also well recognized. Such partial often seen as more natural than a one, even when the symmetric under complex pose. We introduce algorithm to extract reflectional (PIRS) shape. Given closed 2-manifold mesh, we develop voting scheme obtain symmetry axis (IRSA) transform, is scalar field over mesh that accentuates IRSAs then set explicit IRSA curves shape...

Most mesh denoising techniques utilize only either the facet normal field or vertex of a surface. The two fields, though contain some redundant geometry information same model, can provide additional that other lacks. Thus, considering one is likely to overlook geometric features. In this paper, we take advantage piecewise consistent property fields and propose an effective framework in which they are filtered integrated using novel method guide process. Our key observation that, decomposing...

Many geometry processing applications are sensitive to noise and sharp features. Although there a number of works on detecting features in the literature, they heuristic. On one hand, traditional denoising methods use filtering operators remove noise, however, may blur shrink object. other makes detection features, which relies computation differential properties, unreliable unstable. Therefore, discrete surfaces still remains challenging. In this article, we present an approach for...

Abstract We present an adaptive slicing scheme for reducing the manufacturing time 3D printing systems. Based on a new saliency‐based metric, our method optimizes thicknesses of layers to save and preserve visual quality results. formulate problem as constrained ℓ 0 optimization compute result via two‐step scheme. To further reduce time, we develop segmentation partition object into subparts then optimize each subpart separately. validate with large set shapes ranging from CAD models scanned...