Multimodal face aging framework via learning disentangled representation

Existing face aging (FA) approaches usually concentrate on a universal aging pattern, and produce restricted aging faces from one-to-one mapping. However, the diversity of living environments impact individuals differently in their oldness. To simulate various aging effects, we propose a multimodal FA framework based on face disentanglement technique of age-specific and age-irrelevant information. A Variational Autoencoder (VAE)-based encoder is designed to represent the distribution of the age-specific attributes. To capture the age-irrelevant features, a cycle-consistency loss of unpaired faces is utilized among various age spans. The extensive experimental results demonstrate that the sampled age-specific codes along with an age-irrelevant feature make the multimodal FA diverse and realistic.     

Face verification across age progression.

Human faces undergo considerable amounts of varialions with aging. While face recognition systems have been proven to be sensitive to factors such as illumination and pose, their sensitivity to facial aging effects is yet to be studied. How does age progression affect the similarity between a pair of face images of an individual? What is the confidence associated with establishing the identity between a pair of age separated face images? In this paper, we develop a Bayesian age difference classifier that classifies face images of individuals based on age differences and performs face verification across age progression. Further, we study the similarity of faces across age progression. Since age separated face images invariably differ in illumination and pose, we propose preprocessing methods for minimizing such variations. Experimental results using a database comprising of pairs of face images that were retrieved from the passports of 465 individuals are presented. The verification system for faces separated by as many as nine years, attains an equal error rate of 8.5%.     

Facial makeup transfer with GAN for different aging faces

Facial aging is widely used in criminal tracking and the search for lost children. If the aging face is made up, it will greatly affect the discrimination of the tracking system. Therefore, the research on the makeup of different aging faces is extremely important. Existing studies have achieved a good transition from the non-makeup domain to the makeup domain in facial makeup transfer. But few studies involve the transfer of facial makeup at different ages. In addition, existing datasets rarely contain both age and makeup attributes, which make the transfer of facial makeup for different ages full of challenges. To solve the above problems, we propose a learning framework, called AM-Net, which can realize facial makeup transfer for different ages while protecting identity information. AM-Net is composed of two sub-network modules: Aging-Net and Makeup-Net. AM-Net first learns the aging mechanism of faces through Aging-Net, and then, it feeds the learned aging mode to Makeup-Net. After that, AM-Net trains Makeup-Net to realize the mapping relationship between the non-makeup domain to the makeup domain and transfer the makeup style to the face of the non-makeup. Throughout the network, multiple losses are applied to ensure AM-Net preserve information about the identity, background, etc. Extensive experiments are conducted on different datasets with different state-of-the-art methods, which prove the effectiveness of AM-Net.     

An automatic algorithm for semantic object generation and temporal tracking

Automatic semantic video object extraction is an important step for providing content-based video coding, indexing and retrieval. However, it is very difficult to design a generic semantic video object extraction technique, which can provide variant semantic video objects by using the same function. Since the presence and absence of persons in an image sequence provide important clues about video content, automatic face detection and human being generation are very attractive for content-based video database applications. For this reason, we propose a novel face detection and semantic human object generation algorithm. The homogeneous image regions with accurate boundaries are first obtained by integrating the results of color edge detection and region growing procedures. The human faces are detected from these homogeneous image regions by using skin color segmentation and facial filters. These detected faces are then used as object seed for semantic human object generation. The correspondences of the detected faces and semantic human objects along time axis are further exploited by a contour-based temporal tracking procedure.     

3DFP-FCGAN: Face completion generative adversarial network with 3D facial prior

Face completion is a domain-specific image inpainting problem. Most existing face completion methods fail to synthesize fine-grained facial structures due to the undifferentiated treatment of face images and other scene images. To handle this problem, we propose an end-to-end deep generative model based approach which makes full use of the facial prior knowledge, including 2D facial geometry priors from facial parsing maps and landmarks, as well as the 3D depth prior. We adopt a coarse-to-fine inpainting framework where the 2D facial geometry priors based on coarse faces are extracted to guide the refinement network for better planar facial textures and structures. Moreover, a novel 3D regularized reconstruction loss is proposed for the enhancement of the stereo perception of generated faces. Experimental results on two large-scale benchmarks CelebA and CelebA-HQ show that our method significantly outperforms the state-of-the-arts in generating more visually realistic and pleasing faces. Code is available at .     

Adaptive feature selection based on reconstruction residual and accurately located landmarks for expression-robust 3D face recognition

A novel adaptive feature selection based on reconstruction residual and accurately located landmarks for expression-robust 3D face recognition is proposed in this paper. Firstly, the novel facial coarse-to-fine landmarks localization method based on Active Shape Model and Gabor wavelets transformation is proposed to exactly and automatically locate facial landmarks in range image. Secondly, the multi-scale fusion of the pyramid local binary patterns (F-PLBP) based on the irregular segmentation associated with the located landmarks is proposed to extract the discriminative feature. Thirdly, a sparse representation-based classifier based on the adaptive feature selection (A-SRC) using the distribution of the reconstruction residual is presented to select the expression-robust feature and identify the faces. Finally, the experimental evaluation based on FRGC v2.0 indicates that the adaptive feature selection method using F-PLBP combined with the A-SRC can obtain the high recognition accuracy by performing the higher discriminative power to overcome the influence from the facial expression variations.     

Utility Preserved Facial Image De-identification Using Appearance Subspace Decomposition

Automated human facial image de-identification is a much-needed technology for privacy-preserving social media and intelligent surveillance ap-plications. We propose a novel utility preserved facial image de-identification to subtly tinker the appearance of facial images to achieve facial anonymity by creating"averaged identity faces". This approach is able to preserve the utility of the facial images while achieving the goal of privacy protection. We explore a decomposition of an Active appearance model (AAM) face space by using subspace learning where the loss can be modeled as the difference between two trace ratio items, and each respectively models the level of discriminativeness on identity and utility. Finally, the face space is decomposed into subspaces that are respectively sensitive to face identity and face utility. For the subspace most relevant to face identity, a k-anonymity de-identification procedure is applied. To verify the performance of the proposed facial image de-identification approach, we evaluate the created"averaged faces"using the extended Cohn-Kanade Dataset (CK+). The experimental results show that our proposed approach is satisfied to preserve the utility of the original image while defying face identity recognition.     

Enhanced carrier dynamics of PTB7:PC71BM based bulk heterojunction organic solar cells by the incorporation of formic acid

Formic acid (FA) was used as a novel additive in bulk heterojunction (BHJ) solar cells, which contains blends of poly[[4,8-bis[(2-ethylhexyl)oxy]benzo[1,2-b:4,5-b']dithiophene-2,6-diyl][3-fluoro-2-[(2-ethylhexyl)carbonyl]thieno[3,4-b]thiophene-4,6-diyl]] (PTB7) and [6,6]-phenyl-C71-butyric acid methyl ester (PC₇₁BM). The effect of FA on the performance of PTB7:PC₇₁BM based BHJ solar cells is investigated. By the incorporation of FA, the device with the ratio of 6 vol % shows the best power conversion efficiency (PCE) of 9.04%, along with the short-circuit current density (J_sc), open-circuit voltage (V_oc), and fill factor (FF) being 24.11 mA/cm², 0.72 V, and 52.11%, respectively. Experimental results suggest that FA has a strong influence on charge carrier dynamics with a significant increase in J_sc by ∼65% and the dramatically enhanced PCE is mainly due to the increase of absorption and exciton generation of the active layers and the improved charge-carrier mobility of the devices.     

Optimization of slurry components for a copper chemical mechanical polishing at low down pressure using response surface methodology

Copper replaced aluminium and the low-dielectric constant material (low-k dielectrics) served as a better isolator, which has become the inevitable developing trend of IC technology. Due to the low compression resistance of low-k material, the mechanical strength must be reduced in order to ensure the functional integrity, which is a challenge for the traditional chemical-mechanical polishing (CMP) technology. To solve this issue, it is necessary to develop a chemically dominant CMP process at low down pressure. It is generally known that the implementation of optimum slurry composition is one of the important issues. To achieve a high removal rate (RR) and minimal WIWNU (Within-Wafer Non-Uniformity) at a down pressure of 0.63 psi, the response surface methodology (RSM) was applied to optimize slurry composition which contains silica sols, H₂O₂ and FA/O chelating agent. A central composite design, which is the standard design of RSM, was used to evaluate the effects and interactions of three factors. The optimal conditions obtained from the compromise of the two desirable responses, RR and WIWNU, were silica sols concentration of 13.88 vol.%, H₂O₂ concentration of 16.13 ml/L and FA/O chelating agent concentration of 20.22 ml/L, respectively. The RSM was demonstrated as an appropriate approach for the optimization of the slurry components by confirmation experiments.     

Interface Microstructure and Performance of Sb Contacts in Bismuth Telluride-Based Thermoelectric Elements

A thermoelectric joint composed of p-type Bi_0.5Sb_1.5Te₃ (BiSbTe) material and an antimony (Sb) interlayer was fabricated by spark plasma sintering. The reliability of the thermoelectric joints was investigated using electron probe microanalysis for samples with different accelerated isothermal aging time. After aging for 30 days at 300°C in vacuum, the thickness of the diffusion layer at the BiSbTe/Sb interface was about 30 μm, and Sb₂Te₃ was identified to be the major interfacial compound by element analysis. The contact resistivity was 3 × 10^?6 ohm cm² before aging and increased to 8.5 × 10^?6 ohm cm² after aging for 30 days at 300°C, an increase associated with the thickness of the interfacial compound. This contact resistivity is very small compared with that of samples with solder alloys as the interlayer. In addition, we have also investigated the interface behavior of Sb layers integrated with n-type Bi₂Se_0.3Te_2.7 (BiSeTe) material, and obtained similar results as for the p-type semiconductor. The present study suggests that Sb may be useful as a new interlayer material for bismuth telluride-based power generation devices.     

Adult-child 3D backward face aging model (3D B-FAM)

Face aging has been widely considered in many studies regarding all the potential applications. However, the de-aging known as the rejuvenation or backward modeling has recently received more attention. Previous studies mainly focused on rejuvenating faces from aged adults into young adults using two-dimensional (2D) models. In this work, we propose an extension of a previous 2D adult-child B-FAM into 3D model. This model allows a digital face appearance rejuvenation within a range of [75–3] years old. To evaluate the performances of the proposed approach, first, we proposed two performance evaluation modes, namely: Generic Perception Based and Biometric Verification Mode. Then, the performances have been evaluated over our own 3D database, called Face Time-Machine database constructed using 75 females and 70 males, leading to 500 textured surface meshes. Finally, results show that they are perceptually satisfying and system performance increases by using the faces obtained from our model.     

人脸的表情与年龄变换和非完整信息的重构技术(上)

人脸表情变换、年龄变换和非完整信息重构是计算机视觉与图像处理领域的重要研究内容.该文研究了基于线性统计学模型的真实感人像变换和处理技术.应用对人像形状与纹理的矢量化表达,并结合Lambertian表面比率图理论,提出了基于稠密特征对应的人脸复杂表情的合成与克隆算法;根据人像平均化最美的生理学理论,提出了基于平均脸的人像年轻化变换算法,以及基于衰老纹理映射和差异性理论的人像衰老化变换算法.以MPI高加索人像和AIAR东方人像库数据为测试样本的实验结果,验证了人像表情合成、衰老化与年轻化算法的有效性.     

Co-sensitization of N719 and RhCL dyes on carboxylic acid treated TiO2 for enhancement of light harvesting and reduced recombination

We report herein the findings from a systematic study conducted on the effect of co-sensitization of two dyes i.e. N719 and rhodamine 19 perchlorate (RhCL) onto TiO₂ electrodes (thickness: 4 μm), which were treated by three different carboxylic acids namely, formic acid (FA), oxalic acid (OA) and citric acid (CA). Co-sensitization was carried out by first loading the carboxylic acid treated TiO₂ electrodes with RhCL dye followed by N719. The amount of dye loading was estimated from the UV/vis spectroscopy and the results show that: (i) the amount of individual dye loading (RhCL or N719) onto carboxylic acid treated TiO₂ electrodes is, in general, higher than that onto untreated TiO₂ electrodes; (ii) the amount of individual dye loading is highest for OA-treated and lowest for CA-treated TiO₂ electrodes; and (iii) the co-sensitization leads to highest loading of N719 onto FA-treated and lowest on CA-treated TiO₂ electrodes. The results of dye sensitized solar cells fabricated using these acid treated TiO₂ electrodes revealed that the efficiency (η) is higher for electrodes having higher loading of N719 dye. For single N719 dye loading, the highest η of 4.5% is observed for OA-treated TiO₂; while upon co-sensitization the highest η of ∼5.3% is observed for FA treated TiO₂. A detailed analyses of Fourier transform infrared spectroscopy (FTIR), UV–Visible, impendence, incident photon to current efficiency (IPCE) results show that η enhancement occurs due to the following factors: (i) increased short circuit current density (J_SC) owing to high N719 dye loading which enhances light harvesting; (ii) improved IPCE; (iii) increased open circuit voltage (V_OC) due to an upward shift of the conduction band edge (CBE) and quasi Fermi level; and (iv) suppressed back electron transfer.     

Incubating High-Quality Perovskite Film via Facial Mask Technique for Efficient and Stable Solar Cells

High-performance perovskite film with superior internal and surface qualities is critical for perovskite solar cells (PSCs) but hardly achievable due to the rapid crystallization rate of perovskite itself. Herein, a novel technique by in situ manipulating perovskite crystal growth and modifying the surface properties is developed using organic passivating agent-assisted polydimethylsiloxane membrane as a facial mask (FM) of perovskites. By placing the perovskite-precursor films with their faces toward the designed FM during thermal annealing, a favorable microenvironment is constructed for incubating high-quality perovskite films with smooth surface, enhanced vertical orientation of (100) plane, and well-adjusted interfacial energy levels. With this versatile FM incubation technique, efficient PSCs for both methylammonium (MA)-based and formamidinium (FA)-MA-Cs mixed perovskite systems are facilely fabricated, delivering excellent humidity/thermal stabilities and promising efficiencies up to 21.4% with an improved open-circuit voltage of 1.15 V in MA-based devices. This study not only provides a facile and efficient approach to rationally manage the perovskite growth process, but also reveals the fundamental characteristics of high-quality perovskite films comprehensively for the construction of efficient and stable PSCs.     

Comparison of Sn-Ag-Cu Solder Alloy Intermetallic Compound Growth Under Different Thermal Excursions for Fine-Pitch Flip-Chip Assemblies

The intermetallic compound (IMC) evolution in Cu pad/Sn-Ag-Cu solder interface and Sn-Ag-Cu solder/Ni pad interface was investigated using thermal shock experiments with 100-μm-pitch flip-chip assemblies. The experiments show that low standoff height of solder joints and high thermomechanical stress play a great role in the interfacial IMC microstructure evolution under thermal shock, and strong cross-reaction of pad metallurgies is evident in the intermetallic growth. Furthermore, by comparing the IMC growth during thermal aging and thermal shock, it was found that thermal shock accelerates IMC growth and that kinetic models based on thermal aging experiments underpredict IMC growth in thermal shock experiments. Therefore, new diffusion kinetic parameters were determined for the growth of (Cu,Ni)₆Sn₅ using thermal shock experiments, and the Cu diffusion coefficient through the IMC layer was calculated to be 0.2028 μm²/h under thermal shock. Finite-element models also show that the solder stresses are higher under thermal shock, which could explain why the IMC growth is faster and greater under thermal shock cycling as opposed to thermal aging.     

Impact of thermal aging on the intermetallic compound particle size and mechanical properties of lead free solder for green electronics

The Sn-Ag_3.0-Cu_0.5 lead free solder (LFS) alloy is mostly used as good alternative as compared to conventional Tin-Lead (Sn-Pb) due to its good mechanical properties and no harmful effect on environment but it stills has some problems to be solved regarding the growth formation of intermetallic compounds (IMCs). The IMCs present inside the bulk tin (Sn) matrix grow at high temperature and hence their impact on mechanical properties becomes more significant. In this work the effects of lanthanum (La) doping of Sn-Ag_3.0-Cu_0.5 is investigated as function of IMCs growth and mechanical properties including yield strength (YS) and ultimate tensile strength (UTS) under different thermal aging temperatures. The selected La concentration in this study is 0.4 wt%. The aging time is 50 h and thermal aging temperatures are 60 °C, 100 °C and 140 °C. The microstructure examination before and after thermal aging is observed using scanning electron microscopy (SEM) followed by image analysis to estimate the nature of IMCs. The chemical composition is confirmed with energy dispersive X-ray (EDX). The YS and UTS are also examined before and after thermal aging for the un-doped and doped samples from stress-strain curves using universal testing machine (UTM). It is investigated that inclusion of 0.4 wt% La into Sn-Ag_3.0-Cu_0.5 solder system results in increasing the IMCs growth rate and hence mechanical properties reduced. It is also observed that the microstructure becomes coarsen after thermal aging due to growth of average IMCs particle size with significant decrease in YS and UTS. Further, mathematical relations with minimum error are developed to predict mechanical properties (YS and UTS) at various levels of aging temperature, showing a reciprocal relationship between aging temperature and mechanical properties.     

基于Percoll密度梯度法的轻老龄红细胞膜表面电荷的研究

红细胞衰老过程中膜表面电荷的变化情况与其结构功能密切相关。针对以往研究对轻老龄红细胞膜表面电荷的差异性不确定的问题,本文从不同的细胞分离方法分离轻老龄红细胞的效果不同的角度,通过比较分离出的轻老龄红细胞表面唾液酸的含量和细胞膜的弯曲弹性模量Kc值,提出高速离心法和Percoll密度梯度离心法分离出的轻老龄红细胞间的细胞年龄差距不同,Percoll法分离出的细胞年龄差距更大。利用Percoll密度梯度离心法,进一步分离出四种不同细胞年龄的红细胞,由红细胞表面唾液酸的含量和表面Zeta电位值检验轻老龄红细胞膜表面带电状态的差异性,结果显示,轻老龄红细胞的膜表面电荷存在明显差别,且随着细胞年龄的增长,红细胞的膜表面负电荷逐渐减少。研究结果对红细胞的衰老机制及体内清除老化红细胞的途径等研究具有重要意义。     

Examination of a Thermally Viable Structure for an Unconventional Uni-Leg Mg2Si Thermoelectric Power Generator

We have fabricated an unconventional uni-leg structure thermoelectric generator (TEG) element using quad thermoelectric (TE) chips of Sb-doped n-Mg₂Si, which were prepared by a plasma-activated sintering process. The power curve characteristics, the effect of aging up to 500?h, and the thermal gradients at several points on the module were investigated. The observed maximum output power with the heat source at 975?K and the heat sink at 345?K was 341?mW, from which the ??T for the TE chip was calculated to be about 333?K. In aging testing in air ambient, a remarkable feature of the results was that there was no notable change from the initial resistance of the TEG module for as long as 500?h. The thermal distribution for the fabricated uni-leg TEG element was analyzed by finite-element modeling using ANSYS software. To tune the calculation parameters of ANSYS, such as the thermal conductance properties of the corresponding coupled materials in the module, precise measurements of the temperature at various probe points on the module were made. Then, meticulous verification between the measured temperature values and the results calculated by ANSYS was carried out to optimize the parameters.     

Feature Detection and Simplification of 3D Face Data with Facial Expressions

We propose an efficient framework to realistically render 3D faces with a reduced set of points. First, a robust active appearance model is presented to detect facial features in the projected faces under different illumination conditions. Then, an adaptive simplification of 3D faces is proposed to reduce the number of points, yet preserve the detected facial features. Finally, the point model is rendered directly, without such additional processing as parameterization of skin texture. This fully automatic framework is very effective in rendering massive facial data on mobile devices.     

An experimental illustration of 3D facial shape analysis under facial expressions

The main goal of this paper is to illustrate a geometric analysis of 3D facial shapes in the presence of varying facial expressions. This approach consists of the following two main steps: (1) Each facial surface is automatically denoised and preprocessed to result in an indexed collection of facial curves. During this step, one detects the tip of the nose and defines a surface distance function with that tip as the reference point. The level curves of this distance function are the desired facial curves. (2) Comparisons between faces are based on optimal deformations from one to another. This, in turn, is based on optimal deformations of the corresponding facial curves across surfaces under an elastic metric. The experimental results, generated using a subset of the Face Recognition Grand Challenge v2 data set, demonstrate the success of the proposed framework in recognizing people under different facial expressions. The recognition rates obtained here exceed those for a baseline ICP algorithm on the same data set.