2017

• Record 157 of
  
  Title:A novel algorithm for maneuvering target detection under the high energy laser irradiating
  
  Author(s):Ye, Demao(1); Wang, Jing(2); Li, Peizheng(1); Yan, Shiheng(1)
  Source: Proceedings of SPIE - The International Society for Optical Engineering  Volume: 10462  Issue:   DOI: 10.1117/12.2285535  Published: 2017  
  Abstract:The high-energy laser weapon is famous for its unique advantage of speed-of-light response which was considered as an ideal weapon against Unmanned Aerial Vehicle(UAV). However, due to the high energy laser reflection effect, the pixel gray distribution of the frame image will be changed drastically, and therefore the miss distance signal will be interfered strongly when the high energy laser irradiating on the UAV, which seriously affects precision of object tracking in practical application. The traditional "centroid method" or "template matching method" have been difficult to meet the requirements of high precision miss distance which was less than 1pixel(RMS) under the reflected light interfering. In order to developing operational effectiveness of weapon system, G-DS(Gray weighted factor-Diamond Search method) algorithm was proposed which combined with gray weighted factor based on self-learning mechanism. It has been studied for the characteristics of UAV images by field experiment. The results show that G-DS algorithm is low-latency(less than 5ms), which can reduce time complexity compared with the traditional ME algorithm, furthermore, G-DS algorithm was robust based on local motion vector of the block, which can improve ability of target detection and recognition compared with the traditional "centroid method" or "template matching method". Hence, G-DS algorithm was beneficial to the engineering of high-energy laser weapon. © 2017 SPIE.
  Accession Number: 20180404671032
• Record 158 of
  
  Title:Multi-view clustering and semi-supervised classification with adaptive neighbours
  
  Author(s):Nie, Feiping(1); Cai, Guohao(1); Li, Xuelong(2)
  Source: 31st AAAI Conference on Artificial Intelligence, AAAI 2017  Volume:   Issue:   DOI:   Published: 2017  
  Abstract:Due to the efficiency of learning relationships and complex structures hidden in data, graph-oriented methods have been widely investigated and achieve promising performance in multi-view learning. Generally, these learning algorithms construct informative graph for each view or fuse different views to one graph, on which the following procedure are based. However, in many real world dataset, original data always contain noise and outlying entries that result in unreliable and inaccurate graphs, which cannot be ameliorated in the previous methods. In this paper, we propose a novel multi-view learning model which performs clustering/semi-supervised classification and local structure learning simultaneously. The obtained optimal graph can be partitioned into specific clusters directly. Moreover, our model can allocate ideal weight for each view automatically without additional weight and penalty parameters. An efficient algorithm is proposed to optimize this model. Extensive experimental results on different real-world datasets show that the proposed model outperforms other state-of-the-art multi-view algorithms. © Copyright 2017, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved.
  Accession Number: 20174104243241
• Record 159 of
  
  Title:Large-area micro-channel plate photomultiplier tube
  
  Author(s):Sun, Jianning(1); Ren, Ling(1); Cong, Xiaoqing(1); Huang, Guorui(1); Jin, Muchun(1); Li, Dong(1); Liu, Hulin(3); Qiao, Fangjian(1); Qian, Sen(2); Si, Shuguang(1); Tian, Jinshou(2); Wang, Xingchao(1); Wang, Yifang(2); Wei, Yonglin(3); Xin, Liwei(3); Zhang, Haoda(1); Zhao, Tianchi(2)
  Source: Hongwai yu Jiguang Gongcheng/Infrared and Laser Engineering  Volume: 46  Issue: 4  DOI: 10.3788/IRLA201746.0402001  Published: April 25, 2017  
  Abstract:According to the requirement of detector in high energy physics and nuclear physics national scientific equipment, the large-area micro-channel plate photomultiplier(MCP-PMT) different from dynode PMT was researched. The large-area MCP-PMT had low-background glass and microchannel plate multiplier. Using Sb-K-Cs as photocathode, MCP-PMT enjoyed very high quantum efficiency at 350- 450 nm. With double MCPs as electron amplifier, the gain could reach 107. The detection efficiency and single photon detection of large-area PMT was improved. Compared with conventional dynode PMT, this MCP-PMT is a completely new design in structure and has better ratio of spectrum peak to valley, high gain, better anode uniformity, fast response time in single photoelectron detection. © 2017, Editorial Board of Journal of Infrared and Laser Engineering. All right reserved.
  Accession Number: 20172703889299
• Record 160 of
  
  Title:A neighborhood vector principal component analysis method for small defect target detection
  
  Author(s):Wang, Zhengzhou(1,2,3); Yin, Qinye(1); Kou, Jingwei(3); Xia, Yanwen(4); Hu, Bingliang(3)
  Source: Optics InfoBase Conference Papers  Volume: Part F70-PIBM 2017  Issue:   DOI: 10.1364/PIBM.2017.W3A.8  Published: 2017  
  Abstract:The Local Contrast Method (LCM) has many advantages for detecting large defect targets in optical components. However, it often suffers from low performance when the defect target is located in a local bright region, which reduces the accuracy of defect detection. Here, we propose a new Neighborhood Vector Principal Component Analysis (NVPCA) method for small defect target detection. The main idea is that each pixel and its 8 neighbors in the damage image are treated as a column vector for the application of any operations, and a 9-dimensional data cube is reconstructed using the vectors of all pixels. The main information of the data cube is concentrated in the first dimension, therein being the principal component analysis (PCA) transform. When the NVPCA image is again processed using the LCM, a substantial image enhancement is obtained. After extraction of the features of the enhanced image, the important statistical information for each defect target, including coordinates, size, area, and energy integral, can be obtained. Because the defect targets are separated using a region-growing method, this method offers excellent precision in the detection of small defect targets with a size of 1 pixel. In addition, the method can detect defect targets located in local bright regions. © 2017 OSA.
  Accession Number: 20174804476165
• Record 161 of
  
  Title:Modeling Disease Progression via Multisource Multitask Learners: A Case Study with Alzheimer's Disease
  
  Author(s):Nie, Liqiang(1); Zhang, Luming(2); Meng, Lei(3); Song, Xuemeng(4); Chang, Xiaojun(5); Li, Xuelong(6)
  Source: IEEE Transactions on Neural Networks and Learning Systems  Volume: 28  Issue: 7  DOI: 10.1109/TNNLS.2016.2520964  Published: July 2017  
  Abstract:Understanding the progression of chronic diseases can empower the sufferers in taking proactive care. To predict the disease status in the future time points, various machine learning approaches have been proposed. However, a few of them jointly consider the dual heterogeneities of chronic disease progression. In particular, the predicting task at each time point has features from multiple sources, and multiple tasks are related to each other in chronological order. To tackle this problem, we propose a novel and unified scheme to coregularize the prior knowledge of source consistency and temporal smoothness. We theoretically prove that our proposed model is a linear model. Before training our model, we adopt the matrix factorization approach to address the data missing problem. Extensive evaluations on real-world Alzheimer's disease data set have demonstrated the effectiveness and efficiency of our model. It is worth mentioning that our model is generally applicable to a rich range of chronic diseases. © 2012 IEEE.
  Accession Number: 20161002045137
• Record 162 of
  
  Title:Modal simulation and experimental verification of space-borne two dimensional turntable
  
  Author(s):Zou, Dinghua(1,2); Li, Zhiguo(1); Liu, Zhaohui(1); Cui, Kai(1); Zhang, Yongqiang(1,2); Zhou, Liang(1,2)
  Source: Proceedings of SPIE - The International Society for Optical Engineering  Volume: 10463  Issue:   DOI: 10.1117/12.2284587  Published: 2017  
  Abstract:In order to avoid the resonance between the two dimensional turntable and the satellite, the modal simulation of the two dimensional turntable is carried out in this paper. And the simulation results are compared with the experimental results, combined with modal experiment, the simulation results before and after optimization are further verified. Firstly, two dimensional turntable as the research object in this paper, and it is modeled with the finite element method, then we use Patran/Nastran to conduct the modal simulation. In the modal simulation process, the bearing can be equivalent to the spring element, and the MPC element is used to instead of the spring element. And we introduce the modeling method of the MPC unit, the fundamental frequency of two dimensional turntable is obtained through modal simulation. At last, the model experiment is verified by hammering method, the frequency response functions in each direction of x, y and z are measured. Simulations and experimental results show: after optimization, the fundamental frequency of the two dimensional turntable is 42 Hz, which is higher than that of the base frequency 25 Hz, illustrating that the optimized structural design of the two dimensional turntable meets the requirements; The natural frequency and the experimental errors of three-dimensional turntable in x, y, z are 5%, which shows that MPC can simulate the bearing accurately, and is suitable for the simulation of two dimensional turntable. © 2017 SPIE.
  Accession Number: 20180304654855
• Record 163 of
  
  Title:Multifeature anisotropic orthogonal Gaussian process for automatic age estimation
  
  Author(s):Li, Zhifeng(1); Gong, Dihong(2); Zhu, Kai(3); Tao, Dacheng(4,5); Li, Xuelong(6)
  Source: ACM Transactions on Intelligent Systems and Technology  Volume: 9  Issue: 1  DOI: 10.1145/3090311  Published: August 2017  
  Abstract:Automatic age estimation is an important yet challenging problem. It has many promising applications in social media. Of the existing age estimation algorithms, the personalized approaches are among the most popular ones. However, most person-specific approaches rely heavily on the availability of training images across different ages for a single subject, which is usually difficult to satisfy in practical application of age estimation. To address this limitation,we first propose a new model called Orthogonal Gaussian Process (OGP), which is not restricted by the number of training samples per person. In addition, without sacrifice of discriminative power, OGP is much more computationally efficient than the standard Gaussian Process. Based on OGP, we then develop an effective age estimation approach, namely anisotropic OGP (A-OGP), to further reduce the estimation error. A-OGP is based on an anisotropic noise level learning scheme that contributes to better age estimation performance. To finally optimize the performance of age estimation, we propose a multifeature A-OGP fusion framework that uses multiple features combined with a random sampling method in the feature space. Extensive experiments on several public domain face aging datasets (FG-NET, MORPH Album1, and MORPH Album 2) are conducted to demonstrate the state-of-the-art estimation accuracy of our new algorithms. © 2017 ACM.
  Accession Number: 20173904210171
• Record 164 of
  
  Title:On-line dynamic monitoring automotive exhausts: Using BP-ANN for distinguishing multi-components
  
  Author(s):Zhao, Yudi(1,2); Wei, Ruyi(1,2); Liu, Xuebin(1,2)
  Source: Proceedings of SPIE - The International Society for Optical Engineering  Volume: 10461  Issue:   DOI: 10.1117/12.2285325  Published: 2017  
  Abstract:Remote sensing-Fourier Transform infrared spectroscopy (RS-FTIR) is one of the most important technologies in atmospheric pollutant monitoring. It is very appropriate for on-line dynamic remote sensing monitoring of air pollutants, especially for the automotive exhausts. However, their absorption spectra are often seriously overlapped in the atmospheric infrared window bands, i.e. MWIR (3∼5μm). Artificial Neural Network (ANN) is an algorithm based on the theory of the biological neural network, which simplifies the partial differential equation with complex construction. For its preferable performance in nonlinear mapping and fitting, in this paper we utilize Back Propagation-Artificial Neural Network (BP-ANN) to quantitatively analyze the concentrations of four typical industrial automotive exhausts, including CO, NO, NO2 and SO2. We extracted the original data of these automotive exhausts from the HITRAN database, most of which virtually overlapped, and established a mixed multi-component simulation environment. Based on Beer-Lambert Law, concentrations can be retrieved from the absorbance of spectra. Parameters including learning rate, momentum factor, the number of hidden nodes and iterations were obtained when the BP network was trained with 80 groups of input data. By improving these parameters, the network can be optimized to produce necessarily higher precision for the retrieved concentrations. This BP-ANN method proves to be an effective and promising algorithm on dealing with multi-components analysis of automotive exhausts. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.
  Accession Number: 20180404675875
• Record 165 of
  
  Title:Key Fabrication Technology of Polymer Photonic Crystal Fiber for Terahertz Transmission
  
  Author(s):Chen, Qi(1,2); Kong, De-Peng(3); Miao, Jing(3); He, Xiao-Yang(1,2); Zhang, Jian(1,2); Wang, Li-Li(3)
  Source: Guangzi Xuebao/Acta Photonica Sinica  Volume: 46  Issue: 4  DOI: 10.3788/gzxb20174604.0406001  Published: April 1, 2017  
  Abstract:The technologies of fabricating polymer photonics crystal fiber to suit the application needs of terahertz transmission were studied, which were related to material selecting, fiber preform fabrication and fiber drawing. According to the analyzation of optical polymers' properties and the experimental verification, ZEONEX has low absorption of less than 3 cm-1 in Terahertz waves, low water absorption of less than 0.01%, high glass transition tempreture and decomposition temperature of 136℃ and 420℃ respectively. As for fiber preform fabrication and drawing, the model system was improved based on injection moulding, and drawing technology of Pascal level pressure auto-control was initially invented. The controlled value oscillations is no more than 1.5 Pa in the range of 10~200 Pa. Therefore the preform quality and reliability are promoted and fiber microstructure is effectively controlled. With the proposed technology it is hopeful of producing high air filling factor polymer photonics crystal fiber. © 2017, Science Press. All right reserved.
  Accession Number: 20172803903575
• Record 166 of
  
  Title:Window function optimization in atmospheric wind velocity retrieval with doppler difference interference spectrometer
  
  Author(s):Chen, Jiejing(1,2); Feng, Yutao(1); Hu, Bingliang(1); Li, Juan(1); Sun, Jian(1); Hao, Xiongbo(1); Bai, Qinglan(1)
  Source: Guangxue Xuebao/Acta Optica Sinica  Volume: 37  Issue: 2  DOI: 10.3788/AOS201737.0207002  Published: February 10, 2017  
  Abstract:Doppler difference interference spectrometer is a kind of Fourier transform spectrometer. In the process of atmospheric wind velocity retrieval, even-prolongated recovered spectrum cannot work out the phase information of the target spectral line directly. Meanwhile, there are stray spectral lines and noises in the recovered spectrum, which make the phase of the interferogram changed and the retrieved wind velocity deviated. Therefore, isolation of the target spectral line is necessary in the process of getting the phase information of the recovered spectrum in actual noisy environment. For interferograms with different signal noise ratios the retrieved wind velocities (SNR) optimized by different window functions with different line widths are analyzed by Monte-Carlo method. The results indicate that the Gaussian window function with line width equaling 4 to 5 times of the spectral resolution provides the best performance if the SNR of the measured interferogram is higher than 26.5 dB, and rectangular window function with line width equaling 7 to 12 times-of the spectral resolution provides the best performance if the SNR of the measured interferogram is lower than 26.5 dB. The phase information and the approximative atmospheric wind velocity can be retrieved. © 2017, Chinese Lasers Press. All right reserved.
  Accession Number: 20171503569200
• Record 167 of
  
  Title:Identification of isotonic forearm motions using muscle synergies for brain injured patients
  
  Author(s):Geng, Yanjuan(1); Ouyang, Yatao(2); Samuel, Oluwarotimi Williams(1); Yu, Wenlong(1); Wei, Yue(1); Bi, Sheng(3); Lu, Xiaoqiang(4); Li, Guanglin(1)
  Source: International IEEE/EMBS Conference on Neural Engineering, NER  Volume: 0  Issue:   DOI: 10.1109/NER.2017.8008431  Published: August 10, 2017  
  Abstract:To effectively restore the fine motor functions of the forearm and hand of stroke survivors and patients with traumatic brain injury (TBI), recent studies have proposed an active rehabilitation concept based on the pattern recognition of electromyography (EMG) signals to decode the motor intent of the patients. The results from these studies suggested that pattern recognition of EMG signals associated with the limb motions could potentially aid the development of active rehabilitation robots. To obtain richer set of neural information from multiple-channel EMG recordings, this study proposed a muscle synergies based method for motor intent identification from high-density CP EMG signals recorded from eight TBI subjects. For baseline comparison, the linear discriminant analysis (LDA) based pattern recognition approach was also examined. The outcomes show that the proposed muscle synergy based method outperformed the commonly used LDA with more centralized distribution of motion classification accuracy across all the TBI subjects. And such an increment in accuracy suggests the feasibility CP of using muscle synergies for neural control in active rehabilitation for TBI patients. © 2017 IEEE.
  Accession Number: 20173604118932
• Record 168 of
  
  Title:Short-term prediction of UT1-UTC by combination of the grey model and neural networks
  
  Author(s):Lei, Yu(1,2); Guo, Min(3); Hu, Dan-dan(3); Cai, Hong-bing(1,2); Zhao, Dan-ning(1,4); Hu, Zhao-peng(1,4); Gao, Yu-ping(1,2)
  Source: Advances in Space Research  Volume: 59  Issue: 2  DOI: 10.1016/j.asr.2016.10.030  Published: January 15, 2017  
  Abstract:UT1-UTC predictions especially short-term predictions are essential in various fields linked to reference systems such as space navigation and precise orbit determinations of artificial Earth satellites. In this paper, an integrated model combining the grey model GM(1, 1) and neural networks (NN) are proposed for predicting UT1-UTC. In this approach, the effects of the Solid Earth tides and ocean tides together with leap seconds are first removed from observed UT1-UTC data to derive UT1R-TAI. Next the derived UT1R-TAI time-series are de-trended using the GM(1, 1) and then residuals are obtained. Then the residuals are used to train a network. The subsequently predicted residuals are added to the GM(1, 1) to obtain the UT1R-TAI predictions. Finally, the predicted UT1R-TAI are corrected for the tides together with leap seconds to obtain UT1-UTC predictions. The daily values of UT1-UTC between January 7, 2010 and August 6, 2016 from the International Earth Rotation and Reference Systems Service (IERS) 08 C04 series are used for modeling and validation of the proposed model. The results of the predictions up to 30 days in the future are analyzed and compared with those by the GM(1, 1)-only model and combination of the least-squares (LS) extrapolation of the harmonic model including the linear part, annual and semi-annual oscillations and NN. It is found that the proposed model outperforms the other two solutions. In addition, the predictions are compared with those from the Earth Orientation Parameters Prediction Comparison Campaign (EOP PCC) lasting from October 1, 2005 to February 28, 2008. The results show that the prediction accuracy is inferior to that of those methods taking into account atmospheric angular momentum (AAM), i.e., Kalman filter and adaptive transform from AAM to LODR, but noticeably better that of the other existing methods and techniques, e.g., autoregressive filtering and least-squares collocation. © 2016 COSPAR
  Accession Number: 20165203170887