基于单张图像的三维人脸建模与表情动画

在人机交互、数字娱乐等领域,传统的表情合成技术难以稳定地生成具有真实感的个性化人脸表情动画.为此,提出一种基于单张图像的三维人脸建模和表情动画系统.自动检测人脸在图像中的位置,自动定位人脸上的关键点,基于这些关键点和形变模型重建个性化三维人脸模型,对重建的人脸模型进行扩展得到完整的人脸网格,采用稀疏关键点控制的动画数据映射方法来驱动重建的人脸模型生成动态表情动画.实验结果表明,该方法稳定性强、自动化程度高,人脸模型与表情动画比较逼真.     

基于Unity3D的计算机虚拟试衣测量系统设计与实现

在Unity3D引擎环境下,针对虚拟试衣系统中由建模法生成的人体模型难以实现大量个性化定制的问题,提出了一种使用BlendShape算法调整人体各局部特征尺寸,并通过人体特征轮廓曲线约束调整过程的解决方案。该方案由一个基础人体模型向一组极限模型进行过渡变形得到大量个性化人体。为了实现动态试衣,研究了服装基于弹簧-质点模型的布料仿真与碰撞处理过程,由Verlet积分求解的特殊性,通过一组球形碰撞器实现了不同体型的人体模型试穿相同的服装,其结果相较于静态着装避免了穿透现象并提升了真实感。     

人体模型自适应变形算法的研究与实现

针对虚拟试衣个性化调节过程中服装尺寸符合人体模型时出现的人体模型表面穿透服装的失真现象,提出一种人体模型自适应变形算法,通过在需要自适应变形的区域加入质点弹簧系统,并在该系统中加入了一类新的弹簧——支撑弹簧,使得模型表面更接近于真实人体的情况,并对实现算法进行了研究和分析,针对存在的不足进行了改进。实验结果表明,该算法运算效率较高、响应时间短,使用后人体模型能够针对服装的形状进行自适应变形,并具有围度不变性,实现了较为理想的试衣效果。     

基于改进三维形变模型的三维人脸重建和密集人脸对齐方法

针对现在广泛使用的三维形变模型表达能力不够，导致重建出的三维人脸模型泛化性能不佳的问题，提出了一种在姿态、表情和光照未知的条件下的基于单张人脸图片的三维人脸重建和密集人脸对齐的新方法。首先，通过卷积神经网络对现有的三维形变模型进行改进，以提高三维人脸模型的表达能力；然后，基于人脸光滑性和图像相似性，在特征点和像素层面提出新的损失函数，并使用弱监督学习训练卷积神经网络模型；最后，通过训练出的网络模型进行三维人脸重建和密集人脸对齐。实验结果表明，对于三维人脸重建任务，所提模型在AFLW2000-3D上实现了2.25的归一化平均误差；对于密集人脸对齐任务，所提模型在AFLW2000-3D和AFLW-LFPA上分别实现了3.80和3.34的归一化平均误差。与原始使用三维形变模型的方法相比，所提模型在三维人脸重建和密集人脸对齐上的归一化平均误差分别降低了7.4%和7.8%。针对不同光照环境以及角度的人脸图片，该网络模型的重建准确，鲁棒性好，且具有较高的三维人脸重建和密集人脸对齐质量。     

一个实用的人脸定制和表情动画编辑系统

介绍了一个简单而行之有效的特定人脸定制和表情动画的编辑系统，首先给定一个内嵌肌肉向量的一般人脸三维多边形网络模型并提供特定人脸的正侧面正交图象，然后采用snake技术自动适配人脸特征线，变分一般模型定制出三维虚拟人脸；接着再用多分辨率样条技术产生无缝的人脸纹理镶嵌图，进而生成高度真实感的虚拟特定人脸，由内嵌肌肉向量的参数化表示，通过编辑参数能赋予三维虚拟人脸各种丰富的表情。由于变分后人脸模型的拓扑不变，因此可用基于三角化的图象metamorphosis实现不同特定人脸间的三维morph，该系统能在廉价的PC平台上实现，其快速、简单而且具有真实感，具有很大的实用价值。     

基于单张图像的三维人脸重建

利用一种基于法线的模型变形方法,从单张图像重建高质量的三维人脸.利用球谐函数和一个初始参考模型计算得到模型上每个顶点的法线,利用法线使参考模型变形.实验结果表明:提出的算法可以从单幅图像重建具有细节的高质量三维人脸.     

服装CAD中个性化三维人体建模

针对三维服装仿真中对各种不同体态特征的人体模型的需求,给出了一种个性化三维人体建模方法。首先对一系列具有不同特征尺寸的成年女性的人体扫描数据进行简化处理,建立具有一致拓扑的人体模型;然后根据不同人体的对应数据点及其相应的特征尺寸,生成各个简化数据点随特征尺寸变化的规律。利用此变化特性,对参考人体模型进行变形得到新尺寸下的人体模型。该方法已在所开发的三维虚拟试衣系统实现,并取得了较为理想的试验效果。     

虚拟服装展示系统的比较研究

虚拟服装展示是借助立体全息技术、互联网或数字互动大屏幕来展示人体或数字角色“穿上”虚拟服装后的动态视觉效果。随着数字媒体艺术、交互技术和立体显示技术的发展,虚拟服装展示系统被广泛用于立体时装设计、服装表演、娱乐秀、商场试衣以及电视广告等领域。本文分析比较了近年来国内外虚拟服装展示领域的技术发展,并归纳总结出当前虚拟服装展示的5种人机交互模式。作者认为：随着全息技术、体感交互、立体视觉和增强现实等智能环境技术的发展,以“私人订制”为代表的个性化、时尚化和智能化的虚拟试衣体验将成为未来人们服装购买体验中不可或缺的重要环节之一。     

基于改进三维形变模型的三维人脸重建和密集人脸对齐方法

针对现在广泛使用的三维形变模型表达能力不够,导致重建出的三维人脸模型泛化性能不佳的问题,提出了一种在姿态、表情和光照未知的条件下的基于单张人脸图片的三维人脸重建和密集人脸对齐的新方法。首先,通过卷积神经网络对现有的三维形变模型进行改进,以提高三维人脸模型的表达能力;然后,基于人脸光滑性和图像相似性,在特征点和像素层面提出新的损失函数,并使用弱监督学习训练卷积神经网络模型;最后,通过训练出的网络模型进行三维人脸重建和密集人脸对齐。实验结果表明,对于三维人脸重建任务,所提模型在AFLW2000-3D上实现了2.25的归一化平均误差;对于密集人脸对齐任务,所提模型在AFLW2000-3D和AFLW-LFPA上分别实现了3.80和3.34的归一化平均误差。与原始使用三维形变模型的方法相比,所提模型在三维人脸重建和密集人脸对齐上的归一化平均误差分别降低了7.4%和7.8%。针对不同光照环境以及角度的人脸图片,该网络模型的重建准确,鲁棒性好,且具有较高的三维人脸重建和密集人脸对齐质量。     

基于UML的网上Click&Dress系统设计与实现

虚拟试衣系统是一种应用于服装电子商务的实时交互平台。为了满足虚拟试衣系统真实性、实时性、方便性等设计要求,该文使用UML建模技术,结合面对对象思想,设计了基于网络的三维点击-试衣模型体系。     

面向2D虚拟试穿的服装推理变形仿真方法

为了提高当前服装E-Commerce领域2D虚拟试穿展示效果,提出一种基于图像处理技术的服装推理变形仿真方法。采用边缘检测方法提取标准模特图像的特征曲线,检测个性化体型模特图像的体型参数值,获取已变形的各特征区域的外轮廓,进而确定服装试穿素材图像待变形区域外轮廓,对服装试穿素材图像待变形各区域进行插值处理,得到变形后的服装试穿素材图像,从而模拟基于个性化体型的试穿效果。仿真实验结果证明了该方法具有较高的鲁棒性与有效性。     

Efficient 3D reconstruction for face recognition

Face recognition with variant pose, illumination and expression (PIE) is a challenging problem. In this paper, we propose an analysis-by-synthesis framework for face recognition with variant PIE. First, an efficient two-dimensional (2D)-to-three-dimensional (3D) integrated face reconstruction approach is introduced to reconstruct a personalized 3D face model from a single frontal face image with neutral expression and normal illumination. Then, realistic virtual faces with different PIE are synthesized based on the personalized 3D face to characterize the face subspace. Finally, face recognition is conducted based on these representative virtual faces. Compared with other related work, this framework has following advantages: (1) only one single frontal face is required for face recognition, which avoids the burdensome enrollment work; (2) the synthesized face samples provide the capability to conduct recognition under difficult conditions like complex PIE; and (3) compared with other 3D reconstruction approaches, our proposed 2D-to-3D integrated face reconstruction approach is fully automatic and more efficient. The extensive experimental results show that the synthesized virtual faces significantly improve the accuracy of face recognition with changing PIE.     

Fully Convolutional Graph Neural Networks for Parametric Virtual Try-On

We present a learning-based approach for virtual try-on applications based on a fully convolutional graph neural network. In contrast to existing data-driven models, which are trained for a specific garment or mesh topology, our fully convolutional model can cope with a large family of garments, represented as parametric predefined 2D panels with arbitrary mesh topology, including long dresses, shirts, and tight tops. Under the hood, our novel geometric deep learning approach learns to drape 3D garments by decoupling the three different sources of deformations that condition the fit of clothing: garment type, target body shape, and material. Specifically, we first learn a regressor that predicts the 3D drape of the input parametric garment when worn by a mean body shape. Then, after a mesh topology optimization step where we generate a sufficient level of detail for the input garment type, we further deform the mesh to reproduce deformations caused by the target body shape. Finally, we predict fine-scale details such as wrinkles that depend mostly on the garment material. We qualitatively and quantitatively demonstrate that our fully convolutional approach outperforms existing methods in terms of generalization capabilities and memory requirements, and therefore it opens the door to more general learning-based models for virtual try-on applications.     

3D garment positioning using Hermite radial basis functions

Virtual Reality - Fitting an elegant 3D garment model onto a target 3D human model is crucial for garment design industry and virtual try-on systems. The alignment of the garment onto virtual...     

FP-VTON：基于注意力机制的特征保持虚拟试衣网络

随着互联网经济和人工智能技术的飞速发展,越来越多的消费者选择在网上购买衣服,虚拟试衣技术可以为消费者提供方便、快捷的试衣服务,为消费者提供更好的网上购物体验。当前,基于二维图像的虚拟试衣方法可以摒弃三维虚拟试衣所需昂贵的硬件成本和时间代价,但是仍然存在无法有效适应模特的不同体型及大姿态动作的问题,无法充分保留目标服装复杂纹理特征和局部细节特征的问题。为此,提出一种基于注意力机制的特征保持虚拟试衣网络FP-VTON,通过服装变形和服装融合两阶段网络生成虚拟试穿结果。针对传统卷积难以适应非刚性物体大尺寸变形的问题在两阶段网络中引入了捕捉全局特征的特征注意力机制,针对TPS变换翘曲严重的问题提出了服装保真损失函数对网格上点间的距离和斜率进行约束。通过与相关工作的定量和可视化定性实验对比,充分验证了FP-VTON在大姿态形变、复杂纹理服装和特殊体型的情况下可以生成更加逼真的图像,更加有效地保留服装的复杂纹理细节和用户的身份信息。     

Computer aided clothing pattern design with 3D editing and pattern alteration

The traditional apparel product development process is a typical iterative ‘optimization’ process that involves trial-and-error. In order to confirm the design and achieve a satisfactory fit, a number of repeated cycles of sample preparation, trial fitting and pattern alteration must be conducted. The process itself is time-consuming, costly, and dependent on the designer’s skills and experience. In this paper, a novel computer aided design (CAD) solution for virtual try-on, fitting evaluation and style editing is proposed to speed up the clothing design process. A series of new techniques from cross parameterization, geometrical and physical integrated deformation, to novel editing methods are proposed. First, a cross parameterization technique is employed to map clothing pattern pieces on a model surface. The pattern can be precisely positioned to form the initial shape with low distortion. Next, a new deformation method called hybrid pop-up is proposed to approximate the virtual try-on shape. This method is an integration of geometrical reconstruction and physical based simulation. In addition, user interactive operations are introduced for style editing and pattern alteration in both 2D and 3D manners. The standard rules regulating pattern editing in the fashion industry can be incorporated in the system, so that the resulting clothing patterns are suitable for everyday production.     

Automatic pose-independent 3D garment fitting

Given a virtual garment model on a reference human model, we propose an automated 3D garment fitting system that fits the garment model to a target human model. The proposed method can transfer garment models between human models without any user guidance even when the reference and target human models have different poses. Our goal is not to resize or deform the original garment model according to the target human model but to yield realistic fitting results of the given garment on the target human models. Using pose-independent segmentation and cloth simulation, we achieve realistic and automatic fitting results in reasonable running time. Our method can replace the time-consuming manual fitting process that is necessary for many applications that use virtual garments, such as games, animations, CAD tools and online clothing stores.     

DeepGarment : 3D Garment Shape Estimation from a Single Image

3D garment capture is an important component for various applications such as free‐view point video, virtual avatars, online shopping, and virtual cloth fitting. Due to the complexity of the deformations, capturing 3D garment shapes requires controlled and specialized setups. A viable alternative is image‐based garment capture. Capturing 3D garment shapes from a single image, however, is a challenging problem and the current solutions come with assumptions on the lighting, camera calibration, complexity of human or mannequin poses considered, and more importantly a stable physical state for the garment and the underlying human body. In addition, most of the works require manual interaction and exhibit high run‐times. We propose a new technique that overcomes these limitations, making garment shape estimation from an image a practical approach for dynamic garment capture. Starting from synthetic garment shape data generated through physically based simulations from various human bodies in complex poses obtained through Mocap sequences, and rendered under varying camera positions and lighting conditions, our novel method learns a mapping from rendered garment images to the underlying 3D garment model. This is achieved by training Convolutional Neural Networks (CNN‐s) to estimate 3D vertex displacements from a template mesh with a specialized loss function. We illustrate that this technique is able to recover the global shape of dynamic 3D garments from a single image under varying factors such as challenging human poses, self occlusions, various camera poses and lighting conditions, at interactive rates. Improvement is shown if more than one view is integrated. Additionally, we show applications of our method to videos.     

Interactive virtual try-on clothing design systems

Clothing computer design systems include three integrated parts: garment pattern design in 2D/3D, virtual try-on and realistic clothing simulation. Some important results have been obtained in pattern design and clothing simulation since the 1980s. However, in the area of virtual try-on, only limited methods have been proposed which are applicable to some defined garment styles or under restrictive sewing assumptions. This paper presents a series of new techniques from virtually sewing up complex garment patterns on human models to visualizing design effects through physical-based real-time simulation. We first employ an hierarchy of ellipsoids to approximate human models in which the bounding ellipsoids are optimized recursively. We also present a new scheme for including contact friction and resolving collisions. Four types of user interactive operation are introduced to manipulate cloth patterns for pre-positioning, virtual sewing and later obtaining cloth simulation. In the cloth simulation, we propose a simplified cloth dynamic model and an integration scheme to realize a high quality real-time cloth simulation. We demonstrate the robustness of our proposed systems by complex garment style virtual try-on and cloth simulation.     

基于移动平台的三维虚拟试发型系统实现及应用

发型对一个人的形象气质有非常重要的影响,由此所产生的消费也日益增多。消费 者更换发型不适合后无法恢复,因此在换发型和发色之前预览非常必要。已有的发型图册和基于 二维图像的试发应用都存在较为明显的缺陷,不能提供真实的效果。为此,结合虚拟现实技术和 基于单张照片的三维人脸重建算法,实现了一个适用于移动端的三维虚拟试发系统,试发效果更 真实,并可进行多角度观察、更换发色和配饰等。系统在移动平台上进行搭建,通过合理的框架 设计和资源管理进行优化,用户体验和实用性更佳,满足移动用户的需求。