检索结果-南通市图书馆

Adaptive sampling and reconstruction for gradient-domain rendering

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computational Visual Media 2024年第5期10卷 885-902页

作者： Yuzhi Liang Tao Liu Yuchi Huo Rui Wang hujun bao State Key Lab of CAD&CG Zhejiang UniversityHangzhou 310058China College of Transport and Communications Shanghai Maritime UniversityShanghai 201306China

Gradient-domain rendering estimates finite difference gradients of image intensities and reconstructs the final result by solving a screened Poisson problem,which shows improvements over merely sampling pixel *** sampling is another orthogonal research area that focuses on distributing samples adaptively in the primal ***,adaptive sampling in the gradient domain with low sampling budget has been less *** idea is based on the observation that signals in the gradient domain are sparse,which provides more flexibility for adaptive *** propose a deep-learning-based end-to-end sampling and reconstruction framework in gradient-domain rendering,enabling adaptive sampling gradient and the primal maps *** conducted extensive experiments for evaluation and showed that our method produces better reconstruction quality than other methods in the test dataset.

关键词： gradient-domain rendering adaptive rendering Monte Carlo rendering deep learning-based Monte Carlo denoising

State of the Art on Deep Learning-enhanced Rendering Methods

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Machine Intelligence Research 2023年第6期20卷 799-821页

作者： Qi Wang Zhihua Zhong Yuchi Huo hujun bao Rui Wang State Key Laboratory of CAD&CG Zhejiang UniversityHangzhou310058China Zhejiang Laboratory Hangzhou311121China

Photorealistic rendering of the virtual world is an important and classic problem in the field of computer *** the development of GPU hardware and continuous research on computer graphics,representing and rendering virtual scenes has become easier and more ***,there are still unresolved challenges in efficiently rendering global illumination *** the same time,machine learning and computer vision provide real-world image analysis and synthesis methods,which can be exploited by computer graphics rendering *** learning-enhanced rendering combines techniques from deep learning and computer vision into the traditional graphics rendering pipeline to enhance existing rasterization or Monte Carlo integration *** state-of-the-art report summarizes recent studies of deep learning-enhanced rendering in the computer graphics ***,we focus on works of renderers represented using neural networks,whether the scene is represented by neural networks or traditional scene *** works are either for general scenes or specific scenes,which are differentiated by the need to retrain the network for new scenes.

关键词： Neural rendering computer graphics scene representation rendering post-processing

Hybrid3D: learning 3D hybrid features with point clouds and multi-view images for point cloud registration

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2023年第7期66卷 77-93页

作者： Bangbang YANG Zhaoyang HUANG Yijin LI Han ZHOU Hongsheng LI Guofeng ZHANG hujun bao State Key Lab of CAD&CG Zhejiang University The Chinese University of Hong Kong

In recent years, point cloud registration has achieved great success by learning geometric features with deep learning techniques. However, existing approaches that rely on pure geometric context still suffer from sensor noise and geometric ambiguities(e.g., flat or symmetric structure), which limit their robustness to real-world scenes. When 3D point clouds are constructed by RGB-D cameras, we can enhance the learned features with complementary texture information from RGB images. To this end, we propose to learn a 3D hybrid feature that fully exploits the multi-view colored images and point clouds from indoor RGB-D scene scans. Specifically, to address the discrepancy of 2D-3D observations, we design to extract informative 2D features from image planes and take only these features for fusion. Then, we utilize a novel soft-fusion module to associate and fuse hybrid features in a unified space while alleviating the ambiguities of 2D-3D feature binding. Finally, we develop a self-supervised feature scoring module customized for our multi-modal hybrid features, which significantly improves the keypoint selection quality in noisy indoor scene scans. Our method shows competitive registration performance with previous methods on two real-world datasets.

关键词： point cloud registration cross-modal feature fusion multi-view feature fusion computer vision deep learning

Survey and evaluation of monocular visual-inertial SLAM algorithms for augmented reality

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Virtual Reality & Intelligent Hardware 2019年第4期1卷 386-410页

作者： Jinyu LI Bangbang YANG Danpeng CHEN Nan WANG Guofeng ZHANG hujun bao State Key Lab of CAD&CG Zhejiang UniversityHangzhou 310058China SenseTime Research Hangzhou 311215China

Although VSLAM/VISLAM has achieved great success,it is still difficult to quantitatively evaluate the localization results of different kinds of SLAM systems from the aspect of augmented reality due to the lack of an appropriate *** AR applications in practice,a variety of challenging situations(e.g.,fast motion,strong rotation,serious motion blur,dynamic interference)may be easily encountered since a home user may not carefully move the AR device,and the real environment may be quite *** addition,the frequency of camera lost should be minimized and the recovery from the failure status should be fast and accurate for good AR *** SLAM datasets/benchmarks generally only provide the evaluation of pose accuracy and their camera motions are somehow simple and do not fit well the common cases in the mobile AR *** the above motivation,we build a new visual-inertial dataset as well as a series of evaluation criteria for *** also review the existing monocular VSLAM/VISLAM approaches with detailed analyses and ***,we select 8 representative monocular VSLAM/VISLAM approaches/systems and quantitatively evaluate them on our *** dataset,sample code and corresponding evaluation tools are available at the benchmark website http://***/eval-vislam/.

关键词： Visual-inertial SLAM Odometry Tracking Localization Mapping Augmented reality

A causal convolutional neural network for multi-subject motion modeling and generation

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computational Visual Media 2024年第1期10卷 45-59页

作者： Shuaiying Hou Congyi Wang Wenlin Zhuang Yu Chen Yangang Wang hujun bao Jinxiang Chai Weiwei Xu State Key Lab of CAD&CG Zhejiang UniversityHangzhou310058China Xmov Shanghai200030China School of Automation Southeast UniversityNanjing210096China

Inspired by the success of WaveNet in multi-subject speech synthesis,we propose a novel neural network based on causal convolutions for multi-subject motion modeling and *** network can capture the intrinsic characteristics of the motion of different subjects,such as the influence of skeleton scale variation on motion ***,after fine-tuning the network using a small motion dataset for a novel skeleton that is not included in the training dataset,it is able to synthesize high-quality motions with a personalized style for the novel *** experimental results demonstrate that our network can model the intrinsic characteristics of motions well and can be applied to various motion modeling and synthesis tasks.

关键词： deep learning optimization motion generation motion denoising motion control

CPT: a pre-trained unbalanced transformer for both Chinese language understanding and generation

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2024年第5期67卷 43-55页

作者： Yunfan SHAO Zhichao GENG Yitao LIU Junqi DAI Hang YAN Fei YANG Zhe LI hujun bao Xipeng QIU School of Computer Science Fudan University Shanghai Key Laboratory of Intelligent Information Processing Fudan University Zhejiang Lab

In this paper, we take the advantage of previous pre-trained models(PTMs) and propose a novel Chinese pre-trained unbalanced transformer(CPT). Different from previous Chinese PTMs, CPT is designed to utilize the shared knowledge between natural language understanding(NLU) and natural language generation(NLG) to boost the performance. CPT consists of three parts: a shared encoder, an understanding decoder, and a generation decoder. Two specific decoders with a shared encoder are pretrained with masked language modeling(MLM) and denoising auto-encoding(DAE) tasks, *** the partially shared architecture and multi-task pre-training, CPT can(1) learn specific knowledge of both NLU or NLG tasks with two decoders and(2) be fine-tuned flexibly that fully exploits the potential of the model. Moreover, the unbalanced transformer saves the computational and storage cost, which makes CPT competitive and greatly accelerates the inference of text generation. Experimental results on a wide range of Chinese NLU and NLG tasks show the effectiveness of CPT.

关键词： pre-trained model transformer language model generation unified model

同方期刊数据库

Parameterization for fitting triangular mesh

在线全文

学校读者我要写书评

暂无评论

Progress in Natural Science:Materials International 2006年第11期16卷 1214-1221页

作者： LIN Hongwei**, WANG Guojin, LIU Ligang and bao hujun(State Key Laboratory of CAD&CG, Zhejiang University, Hangzhou 310027, China) [a]State Key Laboratory of CAD &CG Zhejiang University Hangzhou China

In recent years, with the development of 3D data acquisition equipments, the study on reverse engineering has become more and more important. However, the existing methods for parameterization can hardly ensure that the parametric domain is rectangular, and the parametric curve grid is regular. In order to overcome these limitations, we present a novel method for parameterization of triangular meshes in this paper. The basic idea is twofold: first, because the isotherms in the steady temperature do not intersect with each other, and are distributed uniformly, no singularity (fold-over) exists in the parameterization; second, a 3D harmonic equation is solved by the finite element method to obtain the steady temperature field on a 2D triangular mesh surface with four boundaries. Therefore, our proposed method avoids the embarrassment that it is impossible to solve the 2D quasi-harmonic equation on the 2D triangular mesh without the parametric values at mesh vertices. Furthermore, the isotherms on the temperature field are taken as a set of iso-parametric curves on the triangular mesh surface. The other set of iso-parametric curves can be obtained by connecting the points with the same chord-length on the isotherms sequentially. The obtained parametric curve grid is regular, and distributed uniformly, and can map the triangular mesh surface to the unit square domain with boundaries of mesh surface to boundaries of parametric domain, which ensures that the triangular mesh surface or point cloud can be fitted with the NURBS surface.

关键词： parameterization regular parametric curve grid mesh fitting reverse engineering.

GPU-based dynamic quad stream for forest rendering

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2010年第8期53卷 1539-1545页

作者： LIU Feng, HUA Wei & bao hujun State Key Lab of Computer Aided Design and Computer Graphics, Zhejiang University, Hangzhou 310058, China 1. State Key Lab of Computer Aided Design and Computer Graphics Zhejiang University Hangzhou 310058 China

Rendering large-scale forest scenes in real-time is a challenging problem, since the geometric complexity is far beyond the capabilities of current graphics hardware. In this paper, we propose a novel approach that adopts a dynamic quad stream residing on GPU buffers for view-dependent forest rendering. The incremental quad stream updating process is well-suited to the massively parallel architecture of the GPU pipeline, and it requires no continuous data exchange between CPU and GPU. The implementation of the algorithm is quite straightforward. Despite its simplicity, the algorithm achieves good performance with low memory, storage, and CPU costs.

关键词： real-time rendering large-scale forest quad stream geometry shader

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

基于自适应遗传算法的手势识别

计算机辅助设计与图形学学报 2007年第8期19卷 1056-1062页

作者：王修晖鲍虎军中国计量学院信息工程学院杭州310018 浙江大学CAD&CG国家重点实验室杭州310058

基于小样本库的手势识别是先进人机交互研究中的一个重要分支.根据Tortoise人手模型训练手势模式库并结合交互者的具体手部特征进行手形训练,生成适用于特定交互者的手势模式库.在交互过程中,根据来自一个或多个同步摄像头的视频信息进... 详细信息

基于小样本库的手势识别是先进人机交互研究中的一个重要分支.根据Tortoise人手模型训练手势模式库并结合交互者的具体手部特征进行手形训练,生成适用于特定交互者的手势模式库.在交互过程中,根据来自一个或多个同步摄像头的视频信息进行基于自适应遗传算法的手势识别.实验结果表明,在环境光照基本稳定的条件下,文中算法可以实现鲁棒的实时手势识别.

关键词：虚拟现实手势交互人手模型手势识别自适应遗传算法

维普期刊数据库