Facial expression recognition of intercepted video sequences based on feature point movement trend and feature block texture variation

Facial Expression Recognition (FER) is an important subject of human–computer interaction and has long been a research area of great interest. Accurate Facial Expression Sequence Interception (FESI) and discriminative expression feature extraction are two enormous challenges for the video-based FER. This paper proposes a framework of FER for the intercepted video sequences by using feature point movement trend and feature block texture variation. Firstly, the feature points are marked by Active Appearance Model (AAM) and the most representative 24 of them are selected. Secondly, facial expression sequence is intercepted from the face video by determining two key frames whose emotional intensities are minimum and maximum, respectively. Thirdly, the trend curve which represents the Euclidean distance variations between any two selected feature points is fitted, and the slopes of specific points on the trend curve are calculated. Finally, combining Slope Set which is composed by the calculated slopes with the proposed Feature Block Texture Difference (FBTD) which refers to the texture variation of facial patch, the final expressional feature are formed and inputted to One-dimensional Convolution Neural Network (1DCNN) for FER. Five experiments are conducted in this research, and three average FER rates 95.2%, 96.5%, and 97% for Beihang University (BHU) facial expression database, MMI facial expression database, and the combination of two databases, respectively, have shown the significant advantages of the proposed method over the existing ones.     

基于多尺度双线性卷积神经网络的多角度下车型精细识别

针对多角度下车辆出现一定的尺度变化和形变导致很难被准确识别的问题，提出基于多尺度双线性卷积神经网络（MS-B-CNN）的车型精细识别模型。首先，对双线性卷积神经网络（B-CNN）算法进行改进，提出MS-B-CNN算法对不同卷积层的特征进行了多尺度融合，以提高特征表达能力；此外，还采用基于中心损失函数与Softmax损失函数联合学习的策略，在Softmax损失函数基础上分别对训练集每个类别在特征空间维护一个类中心，在训练过程中新增加样本时，网络会约束样本的分类中心距离，以提高多角度情况下的车型识别的能力。实验结果显示，该车型识别模型在CompCars数据集上的正确率达到了93.63%，验证了模型在多角度情况下的准确性和鲁棒性。     

A Multiscale-Porous Anion Exchange Membrane for Convenient and Scalable Electrokinetic Concentration of Cationic Species

A device able to electrokinetically concentrate cationic samples has many potential medical and industrial applications, but until now has remained undeveloped due to the lack of a commercial anion-permselective material leading to a prohibitively complex fabrication procedure. Herein, a novel multiscale-porous anion exchange membrane (MP-AEM) that enables the convenient and scalable electrokinetic concentration of cationic species is proposed. A mechanically enhanced multiscale-porous structure with a solid framework is realized by adopting polyester resin as an additive to overcome the intrinsic limitations of the AEM material. The scalable MP-AEM-embedded electrokinetic concentrator is devised based on the peculiar properties of the MP-AEM that for allow both ion and fluid transport. With the MP-AEM, the concentrator is fabricated in a highly streamlined manner consisting only of a simple insertion and assembly. The concentration performance of the MP-AEM-embedded electrokinetic concentrator is demonstrated with a positively charged fluorescent dye and a fluorescein-labeled protein, and the results show enrichment factors of 250 and 500, respectively. The MP-AEM makes cationic electrokinetic concentration more accessible and scalable, thereby enabling further progress in a wide range of fields.     

湿式摩擦副滑摩过程温度场分析及实验验证

湿式摩擦副滑摩过程温度场与应力场相互耦合作用,温度场分布受到多种因素影响,其中压力、旋转速度、润滑流量作为湿式摩擦副工作参数对其温度场的影响尤为显著。在理论分析基础上,采用有限元数值模拟分析与实验研究相结合的方法,对摩擦界面温度场时空分布特性进行研究,同时研究界面温度场在摩擦副工作压力、相对转速和润滑流量作用下的变化规律。研究表明:在对偶钢片和摩擦片近外径侧更易出现高温和应力集中区,且对偶钢片相对于摩擦片更易出现温度和应力分布不均匀情况;温度场中高温集中区与应力场中应力集中区相对应,最大温度随着压力增加、相对转速增大、润滑流量减少而显著上升,该结果得到试验结果的验证。     

视频侦查中人像智能分析应用及算法优化

人像智能分析指的是对视频或录像中的人像进行结构化和可视化分析,对目标人物进行性别、年龄、发型等特征的智能识别,这项技术在视频侦查中有极高的应用价值。人像识别早期的算法是通过人工提取特征,通过学习低级视觉特征来针对不同属性进行分类学习,这种基于传统方法的模型表现常常不尽如人意。在计算机视觉领域,通过海量图像数据学习的神经网络比传统方法有更丰富的信息量和特征可以被提取。文章尝试通过深度学习技术训练神经网络模型对行人进行检测和识别,对于衣着不同的行人进行智能识别,具有更好的鲁棒性,提升了视频人像识别的准确率,拓展了人工智能技术在身份识别领域的应用。     

Tailoring the electron-blocking layer by addition of YSZ to Ba-containing anode for improvement of ceria-based solid oxide fuel cell

The in-situ fabrication of an electron-blocking layer between the Ba-containing anode and the ceria-based electrolyte is an effective approach in suppressing the internal electronic leakage in ceria-based solid oxide fuel cell (SOFC). To improve the thickness of the electron-blocking layer and to research the effect of the layer thickness on the improvement of SOFC, a Ba-containing compound (0.6NiO-0.4BaZr_0.1Ce_0.7Y_0.2O_3-δ) modified by Y stabilized zirconia (YSZ) was employed as a composite anode in this research. SEM analyses demonstrated that the thickness of the interlayer can be simply controlled by regulating the proportion of YSZ at anode. The in-situ formed interlayer in the cell with the anode modified by 20?mol% YSZ possesses a thickness of 0.9?µm which is more suitable for the cell achieving an enhanced performance.     

Novel electrospun polymer electrolytes incorporated with Keggin‐type hetero polyoxometalate fillers as solvent‐free electrolytes for lithium ion batteries

In this study, solvent‐free nanofibrous electrolytes were fabricated through an electrospinning method. Polyethylene oxide (PEO), lithium perchlorate and ethylene carbonate were used as polymer matrix, salt and plasticizer respectively in the electrolyte structures. Keggin‐type hetero polyoxometalate (Cu‐POM@Ru‐rGO, Ni‐POM@Ru‐rGO and Co‐POM@Ru‐rGO (POM, polyoxometalate; rGO, reduced graphene oxide)) nanoparticles were synthesized and inserted into the PEO‐based nanofibrous electrolytes. TEM and SEM analyses were carried out for further evaluation of the synthesized filler structures and the electrospun nanofibre morphologies. The fractions of free ions and crystalline phases of the as‐spun electrolytes were estimated by obtaining Fourier transform infrared and XRD spectra, respectively. The results showed a significant improvement in the ionic conductivity of the nanofibrous electrolytes by increasing filler concentrations. The highest ionic conductivity of 0.28 mS cm^?1 was obtained by the introduction of 0.49 wt% Co‐POM@Ru‐rGO into the electrospun electrolyte at ambient temperature. Compared with solution‐cast polymeric electrolytes, the electrospun electrolytes present superior ionic conductivity. Moreover, the cycle stability of the as‐spun electrolytes was clearly improved by the addition of fillers. Furthermore, the mechanical strength was enhanced with the insertion of 0.07 wt% fillers to the electrospun electrolytes. The results implied that the prepared nanofibres are good candidates as solvent‐free electrolytes for lithium ion batteries. © 2020 Society of Chemical Industry