检索结果-南通市图书馆

三维穿衣人体重建综述——从传统方法到高保真模型

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

中国图象图形学报 2024年第9期29卷 2566-2595页

作者：陈鸿鹄陶云帆张举勇中国科学技术大学安徽省图形计算与感知交互重点实验室合肥230026

三维穿衣人体重建,在计算机图形学和三维视觉领域占有重要地位,广泛应用于多个方向。人体穿衣的多样性和动作的复杂性使得穿衣人体的高保真重建变得极其困难。深度学习技术优化了数据特征提取、隐式几何表示和神经渲染等关键环节,也推... 详细信息

三维穿衣人体重建,在计算机图形学和三维视觉领域占有重要地位,广泛应用于多个方向。人体穿衣的多样性和动作的复杂性使得穿衣人体的高保真重建变得极其困难。深度学习技术优化了数据特征提取、隐式几何表示和神经渲染等关键环节,也推动了高保真穿衣人体重建技术的革命性进步。本文综述了人体重建的基本流程和组成模块,如各类输入数据、人体几何与动作表示、参数化模型以及三维到二维的渲染技术。同时,介绍了公开的穿衣人体数据集,简要回顾了近10年来人体重建算法的快速发展。本文详细探讨了几种主要的重建方法:稠密视角重建、非刚性运动重建(non-rigid structure from motion,NRSFM)、基于像素对齐的隐式几何重建以及生成模型方法。特别是,稠密视角重建能够生成高质量的人体几何,而NRSFM方法减少了对多视角的需求。基于像素对齐的方法重建细节丰富的人体几何,而生成模型方法利用多模态输入信息实现重建。最后总结了现有方法,并展望了未来研究方向,包括实现低成本高保真重建、加速重建过程和增强重建结果的可编辑性,以及在自然环境下进行重建的可能性。本文总结了近年来穿衣人体重建技术的进步,同时指出了未来研究可能集中的方向。

关键词：三维人体重建深度学习参数化模型隐式几何表示非刚性运动重建方法生成模型

在线全文

学校读者我要写书评

暂无评论

多模态数字人建模、合成与驱动综述

中国图象图形学报 2024年第9期29卷 2494-2512页

作者：高玄刘东宇张举勇中国科学技术大学安徽省图形计算与感知交互重点实验室合肥230026

多模态数字人是指具备多模态认知与交互能力,且有类人的思维和行为逻辑的真实自然虚拟人。近年来随着计算机视觉与自然语言处理等领域的交叉融合以及蓬勃发展,相关技术取得显著进步。本文讨论在图形学和视觉领域比较重要的多模态人头动... 详细信息

多模态数字人是指具备多模态认知与交互能力,且有类人的思维和行为逻辑的真实自然虚拟人。近年来随着计算机视觉与自然语言处理等领域的交叉融合以及蓬勃发展,相关技术取得显著进步。本文讨论在图形学和视觉领域比较重要的多模态人头动画、多模态人体动画以及多模态数字人形象构建3个主题,介绍其方法论和代表工作。在多模态人头动画主题下介绍语音驱动人头和表情驱动人头两个问题的相关工作。在多模态人体动画主题下介绍基于循环神经网络(recurrent neural networks,RNN)的、基于Transformer的和基于降噪扩散模型的人体动画生成。在多模态数字人形象构建主题下介绍视觉语言相似性引导的虚拟形象构建、基于多模态降噪扩散模型引导的虚拟形象构建以及三维多模态虚拟人生成模型。本文将相关方向的代表性工作进行介绍和归类,对已有方法进行总结,并展望未来可能的研究方向。

关键词：虚拟数字人建模多模态角色动画多模态生成与编辑神经渲染生成模型神经隐式表示

A Survey of Deep Learning-Based Mesh Processing

在线全文

学校读者我要写书评

暂无评论

Communications in Mathematics and Statistics 2022年第1期10卷 163-194页

作者： He Wang juyong zhang School of Mathematical Sciences University of Science and Technology of ChinaHefei 230026People’s Republic of China

In the past ten years,deep learning technology has achieved a great success in many fields,like computer vision and speech ***,large-scale geometry data become more and more available,and the learned geometry priors have been successfully applied to 3D computer vision and computer graphics *** from the regular representation of images,surface meshes have irregular structures with different vertex numbers and ***,the traditional convolution neural networks used for images cannot be directly used to handle surface meshes,and thus,many methods have been proposed to solve this *** this paper,we provide a comprehensive survey of existing geometric deep learning methods formesh *** first introduce the relevant knowledge and theoretical background of geometric deep learning and some basic mesh data knowledge,including some commonly used mesh ***,we review various deep learning models for mesh data with two different types:graph-based methods and mesh structure-based *** also review the deep learning-based applications for mesh *** the final,we give some potential research directions in this field.

关键词： Geometric deep learning Non-Euclidean space Mesh Convolution Spectral domain Spatial domain

Facial optical flow estimation via neural non-rigid 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

Computational Visual Media 2023年第1期9卷 109-122页

作者： Zhuang Peng Boyi Jiang Haofei Xu Wanquan Feng juyong zhang School of Mathematical Sciences University of Science and Technology of ChinaHefei 230026China

Optical flow estimation in human facial video,which provides 2D correspondences between adjacent frames,is a fundamental pre-processing step for many applications,like facial expression capture and ***,it is quite challenging as human facial images contain large areas of similar textures,rich expressions,and large *** characteristics also result in the scarcity of large,annotated realworld *** propose a robust and accurate method to learn facial optical flow in a self-supervised ***,we utilize various shape priors,including face depth,landmarks,and parsing,to guide the self-supervised learning task via a differentiable nonrigid registration *** experiments demonstrate that our method achieves remarkable improvements for facial optical flow estimation in the presence of significant expressions and large rotations.

关键词： human face optical flow self-supervised non-rigid registration neural networks facial priors

Joint head pose and facial landmark regression from depth images

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 2017年第3期3卷 229-241页

作者： Jie Wang juyong zhang Changwei Luo Falai Chen University of Science and Technology of China HefeiAnhui230026China.

This paper presents a joint head pose and facial landmark regression method with input from depth images for realtime application. Our main contributions are: firstly, a joint optimization method to estimate head pose and facial landmarks, i.e., the pose regression result provides supervised initialization for cascaded facial landmark regression, while the regression result for the facial landmarks can also help to further refine the head pose at each stage. Secondly,we classify the head pose space into 9 sub-spaces, and then use a cascaded random forest with a global shape constraint for training facial landmarks in each specific space. This classification-guided method can effectively handle the problem of large pose changes and ***, we have built a 3D face database containing 73 subjects, each with 14 expressions in various head poses. Experiments on challenging databases show our method achieves state-of-the-art performance on both head pose estimation and facial landmark regression.

关键词： head pose facial landmarks depth images

Real-time face view correction for front-facing cameras

在线全文

学校读者我要写书评

暂无评论

Computational Visual Media 2021年第4期7卷 437-452页

作者： Yudong Guo juyong zhang Yihua Chen Hongrui Cai zhangjin Huang Bailin Deng University of Science and Technology of China Hefei 230026China School of Computer Science and Informatics Cardiff UniversityCardiffWalesUK

Face views are particularly important in person-to-person *** between the camera location and the face orientation can result in undesirable facial appearances of the participants during video *** phenomenon is particularly noticeable when using devices where the frontfacing camera is placed in unconventional locations such as below the display or within the *** this paper,we take a video stream from a single RGB camera as input,and generate a video stream that emulates the view from a virtual camera at a designated *** most challenging issue in this problem is that the corrected view often needs out-of-plane head *** address this challenge,we reconstruct the 3D face shape and re-render it into synthesized frames according to the virtual camera *** output the corrected video stream with natural appearance in real time,we propose several novel techniques including accurate eyebrow reconstruction,high-quality blending between the corrected face image and background,and template-based 3D reconstruction of *** system works well for different lighting conditions and skin tones,and can handle users wearing *** experiments and user studies demonstrate that our method provides high-quality results.

关键词： face view correction 3D face reconstruction deep learning online communication

On SD-oriented mechanism of technological innovation

在线全文

学校读者我要写书评

暂无评论

Ecological Economy 2007年第1期3卷 77-84页

作者： Yang Yuwen zhang juyong School of Economics and Management Dalian Nationalities UniversityDalian 116600P.R.China

If one enterprise just takes technological innovation as the way for economic profit but not the combinationalprofit of economy, society and ecology, it will never survive in this fierce competition market. Based on the relationshipbetween technological innovation and sustainable development, this paper brings forward that technological innovationmust take the sustainable1 development of society as the fundamental goal. Technological innovation should be the realimpulse for the development of society, economy and ecology. On the basis of the analysis on previous technologicalinnovation impetus mechanism, combined with the influential factors, this paper puts forward SD-oriented mechanism oftechnological innovation.

关键词： Sustainable development Technological innovation Core competence

Differentiable Deformation Graph-Based Neural Non-rigid Registration

在线全文

学校读者我要写书评

暂无评论

Communications in Mathematics and Statistics 2023年第1期11卷 151-167页

作者： Wanquan Feng Hongrui Cai Junhui Hou Bailin Deng juyong zhang School of Mathematical Sciences University of Science and Technology of ChinaHefei 230026People’s Republic of China School of Data Science University of Science and Technology of ChinaHefei 230026People’s Republic of China Department of Computer Science City University of Hong KongHong KongPeople’s Republic of China School of Computer Science and Informatics Cardiff UniversityCardiffUK

The traditional pipeline for non-rigid registration is to iteratively update the correspondence and alignment such that the transformed source surface aligns well with the target *** the pipeline,the correspondence construction and iterative manner are key to the results,while existing strategies might result in local *** this paper,we adopt the widely used deformation graph-based representation,while replacing some key modules with neural learning-based ***,we design a neural network to predict the correspondence and its reliability confidence rather than the strategies like nearest neighbor search and pair ***,we adopt the GRU-based recurrent network for iterative refinement,which is more robust than the traditional *** model is trained in a self-supervised manner and thus can be used for arbitrary datasets without *** experiments demonstrate that our proposed method outperforms the state-of-the-art methods by a large margin.

关键词： Differentiable deformation graph Non-rigid registration

Farmers' Willingness-to-Pay for Environment and Human Health and Adoption of IPM

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Population,Resources and Environment 2004年第3期2卷 39-44页

作者： zhang juyong Dalian Nationalities University, Dalian 116600, China Dalian Nationalities University Dalian 116600 China

Although chemical control of pests increases crop production, it brings a lot of damage to environment and human health. There exist a number of alternative methods that are not so harmful to environment and human health. However, whether and how much in extent these technologies adopted are plausible depends on the comparison of benefit-cost between chemical control and the alternative control methods (such as Integrated Pest Management, IPM) and farmers' willingness-to-pay (WTP) for envirorment and human health. Using contingent valuation method (CVM), the author investigates farmers' WTP for environment and human health, recognizes the factors influencing WTP, and accordingly points out the importance of pest control technology extension and government regulation of pesticides.

关键词： contingent valuation method (CVM) environment human health Willingness-to-Pay (WTP) technology adoption