人物肖像画及其卡通动画仿真系统的建立

文章主要目的是利用一台普通的个人电脑和一张人脸的正面照片来自动生成具有写实风格或者夸张效果的人物肖像漫画及其动画.AAM模型是一种统计模型,可以很好地应用于人脸的特征点定位.但是AAM模型在信息的处理上忽略了颜色信息的作用,而肤色信息在人脸检测时往往起着非常大的作用.文中主要采用的方法是先利用肤色信息减小AAM搜索的搜索范围,然后基于这种改进的AAM模型,定位照片上的人脸特征,并以此为基础设计了一个人物肖像画生成系统.该算法对近百幅人脸的正面照片进行实验,实验结果令人满意.该方法能够快速生成特定人脸的肖像画并能较好地应用于卡通动画制作向导、多媒体智能娱乐、游戏和网络即时通讯软件等领域中.     

基于多尺度和多方向特征的人脸超分辨率算法

提出一个基于学习的人脸图像超分辨率算法．该算法采用可操纵金字塔学习人脸图像中的低层次局部特征的空间分布，并结合塔状的父结构和局部最优匹配算法来预测最佳先验模型；然后将先验模型结合到贝叶斯最大后验概率框架中；最后使用最速下降法求出最优的高分辨率人脸图像．实验结果表明，该算法生成的高分辨率人脸图像具有较好的视觉效果．     

基于形变模型的真实感三维人脸动画研究

论文提出了一种新的基于三维人脸形变模型,并兼容于MPEG-4的三维人脸动画模型。采用基于均匀网格重采样的方法建立原型三维人脸之间的对齐,应用MPEG-4中定义的三维人脸动画规则,驱动三维模型自动生成真实感人脸动画。给定一幅人脸图像,三维人脸动画模型可自动重建其真实感的三维人脸,并根据FAP参数驱动模型自动生成人脸动画。     

结合深度图像和强度图像的人脸浅浮雕生成算法

针对单纯基于深度图像压缩方法生成的三维浮雕模型容易存在大量噪声,并且易丢失细节的问题,为了生成特征表达准确且具有艺术美感的人脸浅浮雕,提出一种基于三维深度图像和与之对应的二维强度图像的混合人脸浅浮雕的生成算法.首先对三维深度图像的高度场进行压缩生成浮雕基网格;然后提取二维强度图像的灰度信息、梯度信息和显著度信息,并基于显著度信息对基网格的高度场相应地叠加灰度和梯度信息,以保持脸部五官的细节;最后根据显著度信息使用Laplace算子对网格进行光顺处理,使脸部光滑.实验结果表明,与单纯使用基于深度图像压缩方法相比,文中算法生成的浅浮雕模型可以更好地保持五官的细节特征;与基于图像的浮雕生成方法相比,该算法避免了复杂交互操作,可以更好地保持人脸的整体轮廓和形状.     

基于颜色量化的肖像漫画生成方法研究

目前创意文化已经受到国家和地方政府的关注和大力发展,而人脸漫画也受到了众多爱好者的支持。本文在研究非真实感绘制的基础上,提出了一种基于颜色量化的肖像漫画生成算法。该算法首先通过双边滤波对输入图像进行降噪处理,进而提高图像质量;其次对降噪图像进行颜色量化处理,生成图像的色块效果;然后使用Canny边缘检测算子,提取降噪图像的边缘;最后将色块效果图像与边缘图像进行融合,最终生成肖像漫画效果。实验结果表明,对于给定的输入图像,能够有效地生成具有艺术效果的漫画肖像画。     

基于相机运动子空间的人脸图像非均匀去模糊

由于在相机拍摄的人脸中往往会存在较大的模糊,为了有效去除人脸图像的抖动模糊,提出一种基于非均匀去模糊与人脸特殊属性相结合的人脸图像去模糊方法。首先研究非均匀去模糊的原理,并提出通过约束相机运动子空间的方法来估计出相机与人脸之间的相对运动路径。再根据人脸的特殊属性,通过对清晰人脸训练得到一组清晰的人脸字典,建立人脸的先验知识。最后利用得到的非均匀模糊核和人脸字典对模糊人脸图像进行去卷积。实验结果表明,提出的方法相对于现有的去模糊算法可以得到更清晰的人脸图像,对后续的人脸识别有很大的辅助作用。     

基于DCNN算法的非均匀光照图像人脸识别系统研究

人脸识别系统的人脸图像特征识别失误较高,因此文章设计了基于DCNN算法的非均匀光照图像人脸识别系统。硬件方面设计,采用ARM720T处理器与S3C2440芯片。软件方面建立非均匀光照图像人脸识别功能模块,设计软件整体架构。利用动态卷积神经网络（Dynamic Convolution Neural Network,DCNN）算法,设计人脸光照图像特征的识别程序,在非均匀光照条件下仍能有效地识别出人脸特征。测试结果表明,该能够准确识别出人脸特征,人脸系统识别得更加清晰,能够应用于实际生活中。     

基于模型的头部运动估计和面部图像合成

文中讨论一种基于模型的头部运动估计和面部图像合成方法。首先建立了一个基于人脸几何模型的可变形三维面部模型,此模型可根据不同人脸图像特征修正特定人脸模型。为了使特定人脸模型与特定人脸图像相匹配,需根据变形模型修正人脸模型。文中采用自动调整与人机交互相结合的方法实现特定人脸模型匹配。在调整完模型形状之后,应用３个方向的面部图像进行纹理映射生成不同视点方向的面部图像。应用合成面部图像与输入面部图像最佳匹     

基于Android的三维面部模型自适应调整算法

本文基于Android平台获取二维人物图像,使用face++提供的人脸特征点提取SDK对输入照片的人脸特征点进行自动定位来提取二维人物图像的人脸特征点,根据所提取的特征点使用自适应算法进行调整,生成与二维人物图像相适应的三维面部模型。与一般的调整算法相比,本文针对存在角度的二维人物图像提出生成其对应三维面部模型的自适应调整算法。     

采用几何模型切分的人脸纹理图像生成算法

为了克服传统的照片图像拼接方法中利用特征线进行不同照片之间公共交界线定位不准确的缺点,提出一种“几何模型切分”的人脸纹理图像生成算法。通过对人脸几何模型进行切分,以切分后模型图片的轮廓作为边界线裁剪相应的人脸照片,实现不同照片之间交界线的准确对接,并采用柱面纹理映射方法将生成的纹理图像映射到特定人脸几何模型上。实验结果表明,采用提出的“几何模型切分”算法生成的人脸纹理图像进行纹理映射可以得到较好的真实感三维人脸模型,是一种生成人脸纹理图像的有效方法。     

Automatic eyeglasses removal from face images

In this paper, we present an intelligent image editing and face synthesis system that automatically removes eyeglasses from an input frontal face image. Although conventional image editing tools can be used to remove eyeglasses by pixel-level editing, filling in the deleted eyeglasses region with the right content is a difficult problem. Our approach works at the object level where the eyeglasses are automatically located, removed as one piece, and the void region filled. Our system consists of three parts: eyeglasses detection, eyeglasses localization, and eyeglasses removal. First, an eye region detector, trained offline, is used to approximately locate the region of eyes, thus the region of eyeglasses. A Markov-chain Monte Carlo method is then used to accurately locate key points on the eyeglasses frame by searching for the global optimum of the posterior. Subsequently, a novel sample-based approach is used to synthesize the face image without the eyeglasses. Specifically, we adopt a statistical analysis and synthesis approach to learn the mapping between pairs of face images with and without eyeglasses from a database. Extensive experiments demonstrate that our system effectively removes eyeglasses.     

基于五官特征的人脸剪纸生成方法研究

人脸肖像剪纸应该重现生动的图像细节,为了实现这一目标,提出了一种基于五官特征与图像变形算法的两阶段人脸剪纸合成方法。收集艺术家的人脸剪纸创作,分割五官部位并提取各组件的几何特征,建立数字化五官剪纸数据库。对目标人脸图像进行剪纸合成：在第一阶段,标定目标人脸图像的特征点,分割其五官部位,并提取各部位的几何特征,之后分别计算目标人脸五官与剪纸数据库中各对应组件基于几何特征和形状上下文特征的相似性度量值;通过融合几何特征和形状上下文特征,选择匹配相似度较高的剪纸部位,拼接得到初步的人脸剪纸图。在第二阶段,采用薄板样条（Thin Plate Spline,TPS）变形算法对第一阶段合成的人脸剪纸图进行变形,得到最终的剪纸图像。通过多人视觉测评实验,结果表明运用该方法得到的人脸剪纸图能够达到较为满意的效果。     

用于人像提取及半身像合成的生成对抗网络算法

利用生成对抗网络(generative adversarial network,GAN)进行标准上半身人像的合成,从普通人像照片中截取部分区域得到面部对齐后的标准化上半身合成图像,处理后的标准化人像实现了目标主体与背景的分离,可以有效地优化目标识别和分割算法的结果.图像的合成过程分为2个主要步骤,首先利用图像特征识别人脸并截取头部区域,然后以裁切后的头部区域为中心进行上半身人像的合成,得到人脸特征点及头部区域对齐后的上半身合成图像.该算法可以有效地从背景中分离人像区域,利用合成后的图像进行图像分割和评价,可以避免图像背景对于图像识别主体的干扰.通过自有数据集验证了该算法可以改善分割算法的精确度、召回率和F值,最终合成人脸图像的Facenet平均距离及标准差相比现有的人脸图像正则化算法均有减小,通过在CelebA及LFW等通用数据集上的验证测试,显示出算法具有良好的通用性和适应性,该算法可以广泛适用于人像照片的主体提取和人像合成,作为分割和识别等应用的前置步骤.     

A lighting robust fitting approach of 3D morphable model for face reconstruction

Three-dimensional morphable model (3DMM) is a powerful tool for recovering 3D shape and texture from a single facial image. The success of 3DMM relies on two things: an effective optimization strategy and a realistic approach to synthesizing face images. However, most previous methods have focused on developing an optimization strategy under Phong’s synthesis approach. In this paper, we adopt a more realistic synthesis technique that fully considers illumination and reflectance in the 3DMM fitting process. Using the sphere harmonic illumination model (SHIM), our new synthesis approach can account for more lighting factors than Phong’s model. Spatially varying specular reflectance is also introduced into the synthesis process. Under SHIM, the cost function is nearly linear for all parameters, which simplifies the optimization. We apply our new optimization algorithm to determine the shape and texture parameters simultaneously. The accuracy of the recovered shape and texture can be improved significantly by considering the spatially varying specular reflectance. Hence, our algorithm produces an enhanced shape and texture compared with previous SHIM-based methods that recover shape from feature points. Although we use just a single input image in a profile pose, our approach gives plausible results. Experiments on a well-known image database show that, compared to state-of-the-art methods based on Phong’s model, the proposed approach enhances the robustness of the 3DMM fitting results under extreme lighting and profile pose.     

Automatic portrait oil painter: joint domain stylization for portrait images

Everyone has the dream of being in the center of famous art paintings, admired by numerous future generations. However, the dream came true at a huge cost of the painter’s commission in old days. In our paper, another practical choice is provided for everyone to achieve that dream – an automatic portrait oil painter transferring some artistic styles from one single reference painting. To address this issue, we propose a joint-domain image stylization approach, particularly for portrait oil paintings. From the view of artistic appreciation, we analyze an amount of oil painting art works and summarize three critical factors to depict the figure, i.e. color, structure and texture. Based on this point, we separate and represent an artistic work into these three domains. Then, considering their intrinsic properties and following an art creation route, we propose the corresponding approaches to jointly model and transfer the features in these domains. First, a swatch-based color adjustment is proposed to recolor the tone of the input image based on semantic regions corresponding to the references. Second, the main structures of the input image is maintained by sparse reconstruction. Third, a coarse-to-fine texture synthesis is used to enhance the detail oil painting patterns. Extensive experimental results demonstrate that the proposed method achieves desirable results compared with state-of-the-art methods in not only transferring the styles from references but also keeping consistent contents with the given portrait.     

Expressive line drawings of human faces from range images

We propose a novel technique to extract features from a range image and use them to produce a 3D pen-and-ink style portrait similar to a traditional artistic drawing. Unlike most previous template-based,component-based or example-based face sketching methods,which work from a frontal photograph as input,our system uses a range image as input. Our method runs in real-time for models of moderate complexity,allowing the pose and drawing style to be modified interactively. Portrait drawing in our system makes use of occluding contours and suggestive contours as the most important shape cues. However,current 3D feature line detection methods require a smooth mesh and cannot be reliably applied directly to noisy range images. We thus present an improved silhouette line detection algorithm. Feature edges related to the significant parts of a face are extracted from the range image,connected,and smoothed,allowing us to construct chains of line paths which can then be rendered as desired. We also incorporate various portrait-drawing principles to provide several simple yet effective non-photorealistic portrait renderers such as a pen-and-ink shader,a hatch shader and a sketch shader. These are able to generate various life-like impressions in different styles from a user-chosen viewpoint. To obtain satisfactory results,we refine rendered output by smoothing changes in line thickness and opacity. We are careful to provide appropriate visual cues to enhance the viewer’s comprehension of the human face. Our experimental results demonstrate the robustness and effectiveness of our approach,and further suggest that our approach can be extended to other 3D geometric objects.     

基于语义分割的简洁线条肖像画生成方法

针对目前主流的线条提取算法对于区域对比度不明显的边缘的检测能力较弱,且对于所有区域采用无差别、统一化的处理策略,所生成的线条画往往较复杂,非常不利于机器人机械臂绘图的问题,本文提出了一种基于语义分割的简洁线条肖像画生成方法(concise line portrait generation based on semantic segmentation, CLPG-SS)。首先,对人脸图像进行语义分割,将人脸划分为不同的区域,基于不同区域提取边缘轮廓与五官细节线条,进行边缘切向流优化,从而加强方向信息;在此基础上,利用线条图来生成调和图像,并利用优化后的边缘切向流、人脸语义分割结果以及调和图像,针对不同的分割区域调整线条提取方法的参数,实现对细节无关区域的线条过滤和细节重点区域的线条加强,生成简洁线条肖像画。实验结果表明:本文提出的CLPG-SS方法能够有效提取人脸主轮廓线条,并针对不同区域实现了对细节线条的针对性调节,提高了机器人机械臂的绘制效率。     

A portrait photo-to-tattoo transform based on digital tattooing

Tattooing portraits of loved ones is a popular form of love expression and tribute. Tattooing portraits is complicated and challenging because of detailed facial expressions and unique characters of each person. Currently, it is hard for clients to give clear instructions on tattoo designs to tattooists, because there is no effective way to see a portrait tattoo before putting it on the body. In this paper, an algorithm which transforms a given portrait photo to a portrait tattoo is proposed. It takes a portrait photo, a reference portrait tattoo image, a skin image and a set of parameters as inputs. The portrait photo is the person’s face whom the client wants to put on his/her skin. The reference portrait tattoo image is used to control the color and style of the synthetic portrait tattoo. The skin image is taken from the skin region where the client wants to tattoo. By adjusting the parameters, portrait tattoos with different characteristics can be generated. The proposed algorithm uses a series of tailor-made image processing methods and a digital tattoo needle model to perform digital tattooing on the skin image. Comparing with the state-of-the-art style transfer methods, the proposed algorithm produces more realistic portrait tattoos.

     

Face Hallucination: Theory and Practice

In this paper, we study face hallucination, or synthesizing a high-resolution face image from an input low-resolution image, with the help of a large collection of other high-resolution face images. Our theoretical contribution is a two-step statistical modeling approach that integrates both a global parametric model and a local nonparametric model. At the first step, we derive a global linear model to learn the relationship between the high-resolution face images and their smoothed and down-sampled lower resolution ones. At the second step, we model the residue between an original high-resolution image and the reconstructed high-resolution image after applying the learned linear model by a patch-based non-parametric Markov network to capture the high-frequency content. By integrating both global and local models, we can generate photorealistic face images. A practical contribution is a robust warping algorithm to align the low-resolution face images to obtain good hallucination results. The effectiveness of our approach is demonstrated by extensive experiments generating high-quality hallucinated face images from low-resolution input with no manual alignment.