Robust Minimum Squared Error Classification Algorithm with Applications to Face Recognition

( Zhonghua Liu ),( Chunlei Yang ),( Jiexin Pu ),( Gang Liu ),( Sen Liu ) 한국인터넷정보학회 2016 KSII Transactions on Internet and Information Syst Vol.10 No.1

Although the face almost always has an axisymmetric structure, it is generally not symmetrical image for the face image. However, the mirror image of the face image can reflect possible variation of the poses and illumination opposite to that of the original face image. A robust minimum squared error classification (RMSEC) algorithm is proposed in this paper. Concretely speaking, the original training samples and the mirror images of the original samples are taken to form a new training set, and the generated training set is used to perform the modified minimum sqreared error classification(MMSEC) algorithm. The extensive experiments show that the accuracy rate of the proposed RMSEC is greatly increased, and the the proposed RMSEC is not sensitive to the variations of the parameters.

Noisy label based discriminative least squares regression and its kernel extension for object identification

( Zhonghua Liu ),( Gang Liu ),( Jiexin Pu ),( Shigang Liu ) 한국인터넷정보학회 2017 KSII Transactions on Internet and Information Syst Vol.11 No.5

In most of the existing literature, the definition of the class label has the following characteristics. First, the class label of the samples from the same object has an absolutely fixed value. Second, the difference between class labels of the samples from different objects should be maximized. However, the appearance of a face varies greatly due to the variations of the illumination, pose, and expression. Therefore, the previous definition of class label is not quite reasonable. Inspired by discriminative least squares regression algorithm (DLSR), a noisy label based discriminative least squares regression algorithm (NLDLSR) is presented in this paper. In our algorithm, the maximization difference between the class labels of the samples from different objects should be satisfied. Meanwhile, the class label of the different samples from the same object is allowed to have small difference, which is consistent with the fact that the different samples from the same object have some differences. In addition, the proposed NLDLSR is expanded to the kernel space, and we further propose a novel kernel noisy label based discriminative least squares regression algorithm (KNLDLSR). A large number of experiments show that our proposed algorithms can achieve very good performance.

The nonlinear galloping of iced transmission conductor under uniform and turbulence wind

Zhonghua Liu,Chenhui Ding,Jian Qin,Ying Lei 국제구조공학회 2020 Structural Engineering and Mechanics, An Int'l Jou Vol.75 No.4

The analytical approach for stability and response of iced conductor under uniform wind or turbulent wind is presented in this study. A nonlinear dynamic model is established to describe the motion of iced conductor galloping. In the case of uniform wind, the stability condition is derived by analyzing the eigenvalue associated with linearized matrix; The first order and second order approximation of galloping amplitude are obtained using multi-scale method. However, real wind has random characteristics essentially. To accurately evaluate the performance of the galloping iced conductor, turbulence wind should be described by random processes. In the case of turbulence wind, the Lyapunov exponent is conducted to judge the stability condition; The probability density of displacement is obtained by using the path integral method to predict galloping amplitude. An example is proposed to verify the effectiveness of the previous methods. It is shown that the fluctuating component of wind has little influence on the stability of iced conductor, but it can increase galloping amplitude. The analytical results on stability and response are also verified by numerical time stepping method.

Structure learning of exponential family graphical model with false discovery rate control

Liu Yanhong,Zhang Yuhao,Li Zhonghua 한국통계학회 2023 Journal of the Korean Statistical Society Vol.52 No.3

Probabilistic graphical models enjoy great popularity in a wide range of domains due to their ability to model the conditional dependency relationships among random variables. This paper explores the structure learning for the exponential family graphical model with false discovery rate (FDR) control. Most existing FDR-con-trolled structure learning procedures have been designed for the Gaussian graphical model (GGM). A systematic approach for more general exponential family graphical models is still lacking. In this paper, we introduce a unified procedure to learn the structure of the exponential family graphical model with FDR control utilizing the symmetrized data aggregation (SDA) technique via sample splitting, data screening, and information pooling. We show that our method controls FDR asymptotically under some mild conditions. Extensive simulation results and two real-data examples validate the effectiveness of our method.

Adaptive Fuzzy Finite Time Prescribed Performance Control for Long Stroke Hybrid Robots

Qunpo Liu,Ming Ye,Zhonghua Wu,Xuhui Bu,Naohiko Hanajima 제어·로봇·시스템학회 2023 International Journal of Control, Automation, and Vol.21 No.9

This paper focuses on the problem of finite-time trajectory tracking control of long-stroke hybrid robots with uncertain system parameters and external disturbances. A system filtering prescribed performance backstepping control based on adaptive fuzzy compensation is proposed. For the unknown terms corresponding to the parameter uncertainties and unknown disturbances existing in the system, a set of auxiliary filter variables is introduced in the adaptive fuzzy approximator and an estimation error function is constructed. An adaptive law based on the estimated error function is proposed to adjust the adaptive weight parameters of the fuzzy system. An adaptive fuzzy approximation algorithm based on the estimation error is finally constructed to compensate for the performance loss caused by the unknown term. For the system state convergence speed problem, a performance function that can transform the system tracking error into an unconstrained error is presented, and the unconstrained error is used as the backstepping control variables. To avoid the differential explosion problem, a new set of inverse control variables is defined by combining the filter variables obtained based on the low-pass filter and the joint velocity. Based on the above variables, a system filtering-based prescribed performance backstepping control strategy is proposed. Finally, the semi-global practical finite-time stability of the closed-loop system is proved by the Lyapunov function. Experiments on the MATLAB platform verify the effectiveness of the proposed method.

Synthesis of Composite Carbon-Based Nano-Zero-Valent Iron Micro-Electrolytic Material and Removal of Dichloromethanesulfonic Acid in Drinking Water

Kai Liu,Xiaohang Zhang,Hao Wang,Xuwen He,Zhonghua Shangguan,Yali Shi,Shipeng Yang 성균관대학교(자연과학캠퍼스) 성균나노과학기술원 2019 NANO Vol.14 No.5

Composite carbon-based nano-zero-valent iron materials not only have excellent characteristics of nano materials, but also solve the difficulty of recycling nano materials after use, and overcome the shortcomings of the nano materials. The introduction of copper increases the redox reaction activity of the micro-electrolytic material and expands the applicable pH range of the reaction environment. In the range of pH = 3–9, the removal efficiency of dichloromethanesulfonic acid (Cl2-MSA) was maintained at 95.5–99.7% after 30 min. Continuous dynamic simulation experiments verified that the composite carbon-based nano zero-valent iron material has a stable removal effect on the target pollutants.

A 2 GS/s, 6-bit DAC for UWB Applications In 0.18 ㎛ CMOS Technology

Yi Zhang,Zhonghua Liu,Changchun Zhang,Yufeng Guo,Ying Zhang,Xiaopeng Li,Youtao Zhang,Hao Gao 대한전자공학회 2019 Journal of semiconductor technology and science Vol.19 No.6

To satisfy higher and higher transmission rate and broadband requirement of modern communication, high-speed, low-resolution Digital-to-Analog Converters (DACs) becomes the key element. In this paper, the design of a 2 GS/s, 6 bit DAC without calibration for Ultra-wideband (UWB) applications, is presented. The DAC is based on current steering architecture and is segmented with 4 bit unary and 2 bit binary. To realize larger linear range and fast switching, the source degeneration switch is designed instead of the traditional differential switch. The DAC is designed and taped-out in SMIC 0.18 ㎜ CMOS technology and the area is 975 ㎛´ 775 ㎛. The wafer bonding measurement results shows that the DNL is 0.11 LSB, INL is 0.25 LSB. Under a clock frequency of 2 ㎓, the DAC can achieve a SFDR of 51 ㏈ for input signal of 6㎒, and a SFDR of 32.4 ㏈ for Nyquist input while the power consumption is 79 ㎽.

Synthesis and characterization of Ba(TI1-xZRx)O₃ nanopowders by an aqueous coprecipitation method

Xiangbang Cheng,Zhuo Li,Zhonghua Yao,Yang Liu,Zhiyong Yu,Minghe Cao,Hanxing Liu 한양대학교 세라믹연구소 2008 Journal of Ceramic Processing Research Vol.9 No.6

In this study, single phase Ba(Ti1-xZrx)O3 (0 ≤ × ≤ 0.15) nanopowders have been synthesized successfully by an aqueous coprecipitation method at 85 oC without any further heat treatment. The lattice strain calculated by the Hall-Williamson method indicates that the broadening of the XRD peaks is primarily due to the size of particles not to the lattice strain. The increase of the lattice parameter indicates that Zr ions have entered the lattice and a solid solution was successfully obtained. Furthermore, the particle sizes calculated by the Sherrer formula from the XRD patterns are in agreement with the results of SEM and TEM images. The nanopowders obtained can be used to prepare Ba(Ti1-xZrx)O3 ceramics with a tailored grain size and homogeneous composition, and this is significant to investigate the size effect in Ba(Ti1-xZrx)O3 ceramics and understand the relaxor behavior thoroughly. In this study, single phase Ba(Ti1-xZrx)O3 (0 ≤ × ≤ 0.15) nanopowders have been synthesized successfully by an aqueous coprecipitation method at 85 oC without any further heat treatment. The lattice strain calculated by the Hall-Williamson method indicates that the broadening of the XRD peaks is primarily due to the size of particles not to the lattice strain. The increase of the lattice parameter indicates that Zr ions have entered the lattice and a solid solution was successfully obtained. Furthermore, the particle sizes calculated by the Sherrer formula from the XRD patterns are in agreement with the results of SEM and TEM images. The nanopowders obtained can be used to prepare Ba(Ti1-xZrx)O3 ceramics with a tailored grain size and homogeneous composition, and this is significant to investigate the size effect in Ba(Ti1-xZrx)O3 ceramics and understand the relaxor behavior thoroughly.

Synthesis of novel oleanolic acid and ursolic acid in C-28 position derivatives as potential anticancer agents

Tian Tian,Xinyu Liu,이응석,Jingyang Sun,Zhonghua Feng,Long-Xuan Zhao,Chunhui Zhao 대한약학회 2017 Archives of Pharmacal Research Vol.40 No.4

A series of nitrogen-containing derivatives ofoleanolic acid and ursolic acid were prepared by a modificationat C-28 position via esterification with 2-hydroxyaceticacid followed by amidation with amines, such aspiperazine, N-methylpiperazine, and alkane-1, 2-diamines,alkane-1, 4-diamines, alkane-1, 6-diamines. In vitroantiproliferative activities of the compounds preparedtowards MCF-7, Hela and A549 cell lines were evaluatedby a MTT method to show that OA-5a, OA-5b, OA-5c andUA-5a showed somewhat improved antiproliferativeactivities against MCF-7, Hela and A549 cells comparingto that of the positive control, gefitinib.

Effect of ZnO on the microstructures and piezoelectric properties of K0.5Na0.5NbO3 ceramics

Zhuo Li,Xiangbang Cheng,Zhonghua Yao,Yang Liu,Hanxin Liu,Minghe Cao 한양대학교 세라믹연구소 2009 Journal of Ceramic Processing Research Vol.10 No.5

Lead-free ZnO added K0.5Na0.5NbO3 piezoelectric ceramics have been prepared by a conventional solid state process. The XRD results show that all the K0.5Na0.5NbO3 ceramics with various ZnO contents are single phase with an orthorhombic perovskite structure. The bulk density obtained by the Archimedes method indicates that the addition of ZnO improves the density of K0.5Na0.5NbO3 ceramics under the same sintering conditions. Especially, at a level of 1 mol% ZnO, the sample sintered at 1100 oC for 1h exhibits the highest bulk density (ρ = 4.28 g/㎤) and shows optimal piezoelectric properties, d33 = 135 pC/N and kp = 0.40. Lead-free ZnO added K0.5Na0.5NbO3 piezoelectric ceramics have been prepared by a conventional solid state process. The XRD results show that all the K0.5Na0.5NbO3 ceramics with various ZnO contents are single phase with an orthorhombic perovskite structure. The bulk density obtained by the Archimedes method indicates that the addition of ZnO improves the density of K0.5Na0.5NbO3 ceramics under the same sintering conditions. Especially, at a level of 1 mol% ZnO, the sample sintered at 1100 oC for 1h exhibits the highest bulk density (ρ = 4.28 g/㎤) and shows optimal piezoelectric properties, d33 = 135 pC/N and kp = 0.40.