Ranking with uncertain labels and its applications

The techniques for image analysis and classification generally consider the image sample labels fixed and without uncertainties. The rank regression problem studied in this paper is based on the training samples with uncertain labels, which often is the case for the manual estimated image labels. A core ranking model is designed first as the bilinear fusing of multiple candidate kernels. Then, the parameters for feature selection and kernel selection are learned simultaneously by maximum a posteriori for given samples and uncertain labels. The provable convergency Expectation Maximization (EM) method is used for inferring these parameters in an iterative manner. The effectiveness of the proposed algorithm is finally validated by the extensive experiments on age ranking task and human tracking task. The popular FG-NET and the large scale Yamaha aging database are used for the age estimation experiments, and our algorithm outperforms those state-of-the-art algorithms ever reported by other interrelated literatures significantly. The experiment result of human tracking task also validates its advantage over conventional linear regression algorithm. A short version of this paper appeared in ICME07.     

Modeling and design optimization of switched reluctance machine byboundary element analysis and simulation

Nonlinear boundary element analysis provides a more accurate and detailing tool for the design of switched reluctance machines than conventional equivalent-circuit methods. Design optimization through more detailed analysis and simulation can reduce development and prototyping costs and time to market. Firstly, magnetic field modeling of an industrial switched reluctance machine by the boundary element method is reported in this paper. Secondly, performance prediction and dynamic simulation of motor and control design are presented. Thirdly, magnetic forces that cause noise and vibration are studied, to include the effects of motor and control design variations on noise in the design process. Testing of such a motor in the NEMA 215-Frame size is carried out to verify the accuracy of modeling and simulation     

Hg_（1－x）Cd_xTe的扫描光致发光研究

用傅里叶变换红外扫描光致发光方法研究了Ｈｇ１－ｘＣｄｘＴｅ体单晶样品，该方法可直接得到ＨｇＣｄＴｅ晶片组分的二维平面分布，并可得到辐射复合在复合机制中所占比重的平面分布，以及晶体中非平衡载流子寿命的分布     

光纤智能复合材料

对光纤型智能复合材料的工作原理，性能特点和应用前景作了介绍，并讨论了制造成型的技术问题。     

Prognostic factors of primary liver cancer

To investigate the prognostic factors of primary liver cancer (PLC) and improve the long-term results, 1,248 cases of PLC were analysed. Univariate analysis demonstrated that discovery approach, staging of PLC, original gamma-GPT, resection, radical resection, original AFP, tumor size, tumor number, and tumor capsule have very significant effects on prognosis of PLC (all P < 0.001); cirrhosis, HBsAg, local resection, and tumor embolus in portal vein were also significant difference (all P < 0.05); age, sex, original AFP, hepatitis, and differentiation of PLC cells were no significant difference (all P > 0.05). Multivariate analysis demonstrated that original gamma-GPT, radical resection, tumor size, and tumor number were the most significant prognostic factors (all P < 0.001). Some aspects improving long-term survival were discussed.     

金刚石光电导探测器研制及在软X光测量中的初步应用

介绍了我们自己设计制造的天然Ⅱａ型金刚石光导探测器的结构，原理，制造工艺，实验标定及其在ＩＣＦ实验中软Ｘ光测量中的应用等。     

裂纹梁动态响应有限元分析中的线弹簧模型

本文提出了一种以线弹簧模型为基础来分析裂纹梁动态响应的新数值方法．应用能量原理和断裂力学理论首次建立了线弹簧模型的刚度矩阵，从而确立了一种能使二维裂纹问题转化为一维分析的梁的有限元模型．使用这个模型，研究了不同裂纹长度和裂纹位置对悬臂梁固有频率和振型的影响，并把计算所得到的数值结果与现有的实验数据作了比较．结果表明，当无量纲裂纹长度小于０．６时，两者吻合得非常好；反之，两者之间存在较大的误差．对这种误差产生的原因，本文也作了解释．     

煤气管道膨胀节腐蚀和在线包焊技术应用

简单分析了高炉煤气中Cl元素的存在对不锈钢膨胀节的应力腐蚀问题和防止途径，作为一种临时性的防范措施，简要介绍了煤气膨胀节包焊实施步骤和措施。     

Finite element modelling of skeletal muscles coupled with fatigue

In this study, Hill's muscle theory coupled with fatigue was proposed to describe mechanical behaviours of skeletal muscles. The force developed by a fatigued muscle was described by a muscle fatigue formula which was a time function of the activation α_a and the stretch λ. The modified Hill's muscle theory was hence incorporated into a three-dimensional (3D) finite element model using the PAK finite element code. In this paper, the theoretical derivation of the 3D muscle model was firstly described. After presenting the method of establishing the finite element programme, a case example of studying the mechanical response of a frog gastrocnemius muscle was used to illustrate the applicability of the proposed methodology. The effects of the muscle fatigue on the deformation as well as the stress and strain distribution of the frog muscle subject to a cyclic activation function have been determined. An experiment capturing the real-time shape change of a frog muscle was also conducted to assess the applicability of the proposed method. A comparison between the deformed shapes of the predictive model and the frog muscles was also made. It was shown that the method is capable of providing a reasonable model for describing the mechanical behaviour of skeletal muscles.