基于卷积神经网络的表情识别

在传统静态表情识别研究基础上,提出一种简单的人脸裁剪方法,再用浅层卷积神经网络进一步提取特征并进行表情识别。以CK+和JAFFE为实验数据集,进行预处理效果对比实验、数据增强实验、单种表情识别实验和跨数据集六分类实验。结果表明,针对数据量较少的情况,提出的表情识别方法效果明显且鲁棒性更优。     

Hydrogen related degradation in pipeline steel: A review

To support our increasing energy demand, steel pipelines are deployed in transporting oil and natural gas resources for long distances. However, numerous steel structures experience catastrophic failures due to the evolution of hydrogen from their service environments initiated by corrosion reactions and/or cathodic protection. This process results in deleterious effect on the mechanical strength of these ferrous steel structures and their principal components. The major sources of hydrogen in offshore/subsea pipeline installations are moisture as well as molecular water reduction resulting from cathodic protection. Hydrogen induced cracking comes into effect as a synergy of hydrogen concentration and stress level on susceptible steel materials, leading to severe hydrogen embrittlement (HE) scenarios. This usually manifests in the form of induced-crack episodes, e.g., hydrogen induced cracking (HIC), stress-oriented hydrogen induced cracking (SOHIC) and sulfide stress corrosion cracking (SSCC). In this work, we have outlined sources of hydrogen attack as well as their induced failure mechanisms. Several past and recent studies supporting them have also been highlighted in line with understanding of the effect of hydrogen on pipeline steel failure. Different experimental techniques such as Devanathan–Stachurski method, thermal desorption spectrometry, hydrogen microprint technique, electrochemical impedance spectroscopy and electrochemical noise have proven to be useful in investigating hydrogen damage in pipeline steels. This has also necessitated our coverage of relatively comprehensive assessments of the effect of hydrogen on contemporary high-strength pipeline steel processed by thermomechanical controlled rolling. The effect of HE on cleavage planes and/or grain boundaries has prompted in depth crystallographic texture analysis within this work as a very important parameter influencing the corrosion behavior of pipeline steels. More information regarding microstructure and grain boundary interaction effects have been presented as well as the mechanisms of crack interaction with microstructure. Since hydrogen degradation is accompanied by other corrosion-related causes, this review also addresses key corrosion causes affecting offshore pipeline structures fabricated from steel. We have enlisted and extensively discussed several recent corrosion mitigation trials and performance tests in various media at different thermal and pressure conditions.     

Highly textured and strongly anisotropic TiB2 ceramics prepared using magnetic field alignment (9T)

Highly textured TiB₂ ceramics were prepared by slip casting an aqueous suspension in a magnetic field of 9 T, followed by sintering using Field Assisted Sintering Technology (FAST). Particle size refinement by ball milling improved both the degree of texturing and densification of the material (RD > 98 %). The sintered material exhibited a Lotgering orientation factor of 0.90, with the c-axis of TiB₂ oriented parallel to the magnetic field and FAST pressing direction. The texturing effect induced by the uniaxial pressing was negligible. The textured TiB₂ material exhibited a significant anisotropy in mechanical properties; the values of hardness and indentation elastic modulus measured along directions transverse to the c-axis of TiB₂ were 37 % and 13 % higher than the ones measured along the c-axis direction. Moreover, the specific wear rate of a surface of textured TiB₂ parallel to the field was one order of magnitude lower than a surface perpendicular to the field.     

一种广义不可分支持向量机算法

针对标准的C-SVM(C-support vector machine)算法在处理很多实际分类问题时，对识别错误代价损失差异很大的极端情况表现出的局限性，提出一种通用的 广义支持向量机算法。根据识别错误后所付出的代价，可以把最优分类面向代价损失低的一方进行推移，留给代价损失高的一方更大的空间，提高其识别率，从而减小识别错误后带来的代价损失。该方法进一步提高了标准C SVM的适用性以及样本的正确识别率，将新算法应用到高分辨雷达距离像的识别中，实验证明,广义C-SVM能取得比传统C-SVM更好的识别效果。     

Rapidly finding CAD features using database optimization

Automatic feature recognition aids downstream processes such as engineering analysis and manufacturing planning. Not all features can be defined in advance; a declarative approach allows engineers to specify new features without having to design algorithms to find them. Naive translation of declarations leads to executable algorithms with high time complexity. Database queries are also expressed declaratively; there is a large literature on optimizing query plans for efficient execution of database queries. Our earlier work investigated applying such technology to feature recognition, using a testbed interfacing a database system (SQLite) to a CAD modeler (CADfix). Feature declarations were translated into SQL queries which are then executed.The current paper extends this approach, using the PostgreSQL database, and provides several new insights: (i) query optimization works quite differently in these two databases, (ii) with care, an approach to query translation can be devised that works well for both databases, and (iii) when finding various simple common features, linear time performance can be achieved with respect to model size, with acceptable times for real industrial models. Further results also show how (i) lazy evaluation can be used to reduce the work performed by the CAD modeler, and (ii) estimating the time taken to compute various geometric operations can further improve the query plan. Experimental results are presented to validate our main conclusions.     

Effect of the interfacial condition on the microtexture near the interface of Al/Cu composites during multi-pass caliber rolling

The local microtexture developments of Cu near the interface of Al-core/Cu-sheath composites during multi-pass caliber rolling are investigated using the finite element method in conjunction with a visco-plastic self-consistent model. Two models with different interfacial conditions between Al and Cu are used in order to investigate the effect of the interfacial condition. The resulting inverse pole figures and difference in ODFs for the different numbers of caliber rolling passes indicate that the Coulomb frictional contact condition between Al and Cu represents the final microtextures better than the fully bonded interfacial condition.