2016

• Record 133 of
  
  Title:Salient Band Selection for Hyperspectral Image Classification via Manifold Ranking
  
  Author(s):Wang, Qi(1); Lin, Jianzhe(2); Yuan, Yuan(2)
  Source: IEEE Transactions on Neural Networks and Learning Systems  Volume: 27  Issue: 6  DOI: 10.1109/TNNLS.2015.2477537  Published: June 2016  
  Abstract:Saliency detection has been a hot topic in recent years, and many efforts have been devoted in this area. Unfortunately, the results of saliency detection can hardly be utilized in general applications. The primary reason, we think, is unspecific definition of salient objects, which makes that the previously published methods cannot extend to practical applications. To solve this problem, we claim that saliency should be defined in a context and the salient band selection in hyperspectral image (HSI) is introduced as an example. Unfortunately, the traditional salient band selection methods suffer from the problem of inappropriate measurement of band difference. To tackle this problem, we propose to eliminate the drawbacks of traditional salient band selection methods by manifold ranking. It puts the band vectors in the more accurate manifold space and treats the saliency problem from a novel ranking perspective, which is considered to be the main contributions of this paper. To justify the effectiveness of the proposed method, experiments are conducted on three HSIs, and our method is compared with the six existing competitors. Results show that the proposed method is very effective and can achieve the best performance among the competitors. © 2012 IEEE.
  Accession Number: 20161302165231
• Record 134 of
  
  Title:Surface ship target detection in hyperspectral images based on improved variance minimum algorithm
  
  Author(s):Wang, Zhengzhou(1,2,3); Yin, Qinye(2); Li, Hongguang(3); Hu, Bingliang(3)
  Source: Proceedings of SPIE - The International Society for Optical Engineering  Volume: 10033  Issue:   DOI: 10.1117/12.2243872  Published: 2016  
  Abstract:In order to realize the effective detection of surface structure targets in hyperspectral images, an improved target detection algorithm was proposed in this paper presents to solve the CEM algorithm problems which the large object extraction efficiency is low.First, the image was preprocessed, including end-member extraction, SAM classification. Second, after the ship pixels were subtracted from all pixels, the correlation matrix of pure background pixels was constructed to detect ship target. Third, the biggest write region was found as sea region by mathematical morphology. Finally, the false target pixels were removed from all target pixels using the characteristics which ship targets were surrounded in seawater, so the final ship targets were selected in the end. Experimental results show that the final max ratio between the energy of detection target and the energy of background increased greatly, the target signal is enhanced and the background signal is suppressed by the improved algorithm. © 2016 SPIE.
  Accession Number: 20164903101484
• Record 135 of
  
  Title:Parameter-free auto-weighted multiple graph learning: A framework for multiview clustering and semi-supervised classification
  
  Author(s):Nie, Feiping(1); Li, Jing(1); Li, Xuelong(2)
  Source: IJCAI International Joint Conference on Artificial Intelligence  Volume: 2016-January  Issue:   DOI:   Published: 2016  
  Abstract:Graph-based approaches have been successful in unsupervised and semi-supervised learning. In this paper, we focus on the real-world applications where the same instance can be represented by multiple heterogeneous features. The key point of utilizing the graph-based knowledge to deal with this kind of data is to reasonably integrate the different representations and obtain the most consistent manifold with the real data distributions. In this paper, we propose a novel framework via the reformulation of the standard spectral learning model, which can be used for multiview clustering and semisupervised tasks. Unlike other methods in the literature, the proposed methods can learn an optimal weight for each graph automatically without introducing an additive parameter as previous methods do. Furthermore, our objective under semisupervised learning is convex and the global optimal result will be obtained. Extensive empirical results on different real-world data sets demonstrate that the proposed methods achieve comparable performance with the state-of-the-art approaches and can be used more practically.
  Accession Number: 20165103147035
• Record 136 of
  
  Title:Design of thermal vacuum quasi-accelerated life test for space two-axis gimbal
  
  Author(s):Shangguan, Ai-Hong(1,2); Zhang, Hao-Su(1,2); Wang, Chen-Jie(1,3); Qin, De-Jin(1); Liu, Zhao-Hui(1)
  Source: Jilin Daxue Xuebao (Gongxueban)/Journal of Jilin University (Engineering and Technology Edition)  Volume: 46  Issue: 1  DOI: 10.13229/j.cnki.jdxbgxb201601028  Published: January 1, 2016  
  Abstract:The domestic and abroad thermal vacuum quasi-accelerated life tests of two-axis gimbals are investigated. It is shown that almost all the tests are focused on the parts rather than on a full gimbal, which can not reflect the performance of the full gimbal. According to the analysis of the structure and lubrication mode of the thermal vacuum quasi-accelerated life tests of two-axis gimbals, the five main elements and their design methods are determined. The test contents and methods are designed for the motion performance examination of a full gimbal, and the failure criterion is given to control the test process. A thermal vacuum quasi-accelerated life test is designed in accordance with the design method of the element design method. A series of test data about the motion performance of the two-axis gimbal is acquired by experiments, which can evaluate the in-orbit life of the space two-axis gimbal. The in-orbit application of the obtained results demonstrate that the designed thermal vacuum quasi-accelerated life test method for the space two-axis gimbal can satisfy the test requirement and it is effective and feasible. © 2016, Editorial Board of Jilin University. All right reserved.
  Accession Number: 20160701935595
• Record 137 of
  
  Title:Temporal Multimodal Learning in Audiovisual Speech Recognition
  
  Author(s):Hu, Di(1); Li, Xuelong(2); Lu, Xiaoqiang(2)
  Source: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition  Volume: 2016-December  Issue:   DOI: 10.1109/CVPR.2016.389  Published: December 9, 2016  
  Abstract:In view of the advantages of deep networks in producing useful representation, the generated features of different modality data (such as image, audio) can be jointly learned using Multimodal Restricted Boltzmann Machines (MRB-M). Recently, audiovisual speech recognition based the M-RBM has attracted much attention, and the MRBM shows its effectiveness in learning the joint representation across audiovisual modalities. However, the built networks have weakness in modeling the multimodal sequence which is the natural property of speech signal. In this paper, we will introduce a novel temporal multimodal deep learning architecture, named as Recurrent Temporal Multimodal RB-M (RTMRBM), that models multimodal sequences by transforming the sequence of connected MRBMs into a probabilistic series model. Compared with existing multimodal networks, it's simple and efficient in learning temporal joint representation. We evaluate our model on audiovisual speech datasets, two public (AVLetters and AVLetters2) and one self-build. The experimental results demonstrate that our approach can obviously improve the accuracy of recognition compared with standard MRBM and the temporal model based on conditional RBM. In addition, RTMRBM still outperforms non-temporal multimodal deep networks in the presence of the weakness of long-term dependencies. © 2016 IEEE.
  Accession Number: 20170403274625
• Record 138 of
  
  Title:A Local Structural Descriptor for Image Matching via Normalized Graph Laplacian Embedding
  
  Author(s):Tang, Jun(1); Shao, Ling(2); Li, Xuelong(3); Lu, Ke(4,5)
  Source: IEEE Transactions on Cybernetics  Volume: 46  Issue: 2  DOI: 10.1109/TCYB.2015.2402751  Published: February 2016  
  Abstract:This paper investigates graph spectral approaches to the problem of point pattern matching. Specifically, we concentrate on the issue of how to effectively use graph spectral properties to characterize point patterns in the presence of positional jitter and outliers. A novel local spectral descriptor is proposed to represent the attribute domain of feature points. For a point in a given point-set, weight graphs are constructed on its neighboring points and then their normalized Laplacian matrices are computed. According to the known spectral radius of the normalized Laplacian matrix, the distribution of the eigenvalues of these normalized Laplacian matrices is summarized as a histogram to form a descriptor. The proposed spectral descriptor is finally combined with the approximate distance order for recovering correspondences between point-sets. Extensive experiments demonstrate the effectiveness of the proposed approach and its superiority to the existing methods. © 2013 IEEE.
  Accession Number: 20151200650993
• Record 139 of
  
  Title:Analysis on deployment error of segmented mirror optical system
  
  Author(s):Wang, Chenchen(1,2); Zou, Gangyi(1,2); Pang, Zhihai(1); Li, Ruichang(1,2); Fan, Xuewu(1)
  Source: Guangxue Xuebao/Acta Optica Sinica  Volume: 36  Issue: 11  DOI: 10.3788/AOS201636.1122005  Published: November 10, 2016  
  Abstract:The main method to improve the resolution of optical system is enlarging the pupil of the optical system, and segmenting sub-mirrors to get an equivalent large diameter primary mirror is a common way to enlarge clear aperture. After the deployment of sub-mirrors for segmented mirrors optical system on orbit, there is deviation between deployment position and the designed position, which is called position error. The error determines the imaging quality of the optical system. So it is necessary to analyze the precision of the position of sub-mirrors. Optical software is used to model the segmented mirror optical system, and the relationship curves between position error and the system's imaging quality is got by adjusting position error of six degrees of freedom for sub-mirrors. The results show that root mean square (RMS) values generated by the same position error are different when sub-mirrors are at the different positions. The middle mirrors are sensitive to the movement along X axis direction, and the outer ones are sensitive to the movement along the Y axis direction. Two methods are used to distribute the errors into the sub-mirror position error. One is distributing the position error to each sub-mirror alone. Another is distributing the same position error to each sub-mirror. The results show that the former way has a more relax position error when the generated wavefront errors are the same. © 2016, Chinese Lasers Press. All right reserved.
  Accession Number: 20165003118622
• Record 140 of
  
  Title:Influence of three-point-mount-induced deformation for a three-mirror-anastigmatic telescope on its astigmatic field
  
  Author(s):Zou, Gangyi(1,2); Fan, Xuewu(1); Pang, Zhihai(1); Wang, Chenchen(1,2); Li, Ruichang(1,2)
  Source: Guangxue Xuebao/Acta Optica Sinica  Volume: 36  Issue: 6  DOI: 10.3788/AOS201636.0622002  Published: June 10, 2016  
  Abstract:In order to improve the efficiency of alignment for a three-mirror-anastigmatic (TMA) telescope with large aperture, the influence of three point mount induced (TPMI) deformation on its astigmatic field distribution in the aligning process is analyzed. Based on the vector aberration theory and aperture coordinate transformation, the astigmatic field distribution characters for the aligned and misaligned TMA telescope are analyzed with TPMI deformation on mirrors located at or away from the aperture stop. Finally, the surface error due to TPMI deformation is simulated by the optical design software CODE V using Z10 and Z11, the terms of fringe Zernike polynomial. The characteristic field distribution of astigmatism is verified by the real-ray-tracing method. The results indicate that the astigmatic field distributions are not affected when the TPMI deformation is on the primary mirror (aperture stop), and it shows different characteristic field distributions when the TPMI deformation is on the secondary or tertiary mirror (not aperture stop) for the aligned or misaligned telescope because of producing astigmatism with the field conjugated and field linear. The aligning state of TMA telescope can be qualitatively analyzed by its astigmatic field distribution at the final aligning stage, which is quite instructive for the alignment of TMA telescopes with large aperture. © 2016, Chinese Lasers Press. All right reserved.
  Accession Number: 20162502529806
• Record 141 of
  
  Title:Effects of Laser Repetition Rate and Fluence on Micromachining of TiC Ceramic
  
  Author(s):Zhang, Ying(1); Wang, Yuqian(1); Zhang, Junzhan(1); Liu, Yongsheng(2); Yang, Xiaojun(3); Li, Weinan(3)
  Source: Materials and Manufacturing Processes  Volume: 31  Issue: 7  DOI: 10.1080/10426914.2015.1037916  Published: May 18, 2016  
  Abstract:Micromachining of titanium carbide (TiC) ceramic is very difficult because of its high hardness and brittleness. Femtosecond pulsed laser was employed to process circular rings on the surface of TiC ceramic. The interaction area between femtosecond laser pulses and TiC at different laser repetition rates and fluences was studied. Morphology and composition of irradiated area were analyzed by scanning electron microscope, energy dispersive spectroscopy, and Raman spectrum. The results indicated that the radius of outer circle was close to the intended radius. Laser fluence had obvious effects on the radius and width of circular rings, compared to laser repetition rate. The width of circular rings increased rapidly with increasing laser fluence from 2.55 × 10-2 to 1.27 × 10-1 J/mm2, and then stabilized at around 40 μm when laser fluence was above 7.64 × 10-1 J/mm2. The surface of circular rings was characterized by ripples at the lower laser fluence. With increasing laser fluence, four kinds of typical morphology were observed, including ripples, cauliflower-like particles, ball-like particles, and deposited oxide layer. Ball-like particles contained high concentration of titanium, which came from melt ball splashing from ablation area. The others came from the different oxidation stages occurred on the surface of TiC sample. © Copyright 2016 Taylor & Francis Group, LLC.
  Accession Number: 20153801286085
• Record 142 of
  
  Title:Material research on the InGaAs-emitting-layer VECSEL grown on GaAs substrate
  
  Author(s):Lin, Tao(1); Sun, Ruijuan(1); Sun, Hang(1); Guo, Enmin(1); Duan, Yupeng(2); Lin, Nan(3); Ma, Xiaoyu(3); Wang, Yonggang(4)
  Source: Materials Science in Semiconductor Processing  Volume: 42  Issue:   DOI: 10.1016/j.mssp.2015.11.002  Published: February 1, 2016  
  Abstract:Wafers of InGaAs-emitting-layer vertical external cavity surface emitting semiconductor laser (VECSEL) gain chip and separate active region were grown on semi-insulator GaAs substrates by low pressure metal-organic vapor phase epitaxy (MOVPE). Photoluminescence (PL) wavelength of the active region could be adjusted linearly about 1 nm for increasing 1 sccm H2 flow rate through TMIn under AsH3 flow rates of 150 sccm. The complicated surface-emitted PL signal of the VECSEL gain chip was strongly modulated by interferences within the multilayer and was interpreted by the aberrance of the quantum wells emission with a profile filtered by a micro-cavity resonance in the longitudinal confinement factor. Material tests of the VECSEL wafer showed the reflectivity of the DBR mirrors was in good agreement with the active region photoluminescence, and the wafer was obtained with high crystal quality. © 2015 Elsevier Ltd. All rights reserved.
  Accession Number: 20155001651058
• Record 143 of
  
  Title:Learning spatio-temporal representations for action recognition: A genetic programming approach
  
  Author(s):Liu, Li(1,2); Shao, Ling(1,2); Li, Xuelong(3); Lu, Ke(4,5)
  Source: IEEE Transactions on Cybernetics  Volume: 46  Issue: 1  DOI: 10.1109/TCYB.2015.2399172  Published: January 2016  
  Abstract:Extracting discriminative and robust features from video sequences is the first and most critical step in human action recognition. In this paper, instead of using handcrafted features, we automatically learn spatio-temporal motion features for action recognition. This is achieved via an evolutionary method, i.e., genetic programming (GP), which evolves the motion feature descriptor on a population of primitive 3D operators (e.g., 3D-Gabor and wavelet). In this way, the scale and shift invariant features can be effectively extracted from both color and optical flow sequences. We intend to learn data adaptive descriptors for different datasets with multiple layers, which makes fully use of the knowledge to mimic the physical structure of the human visual cortex for action recognition and simultaneously reduce the GP searching space to effectively accelerate the convergence of optimal solutions. In our evolutionary architecture, the average cross-validation classification error, which is calculated by an support-vector-machine classifier on the training set, is adopted as the evaluation criterion for the GP fitness function. After the entire evolution procedure finishes, the best-so-far solution selected by GP is regarded as the (near-)optimal action descriptor obtained. The GP-evolving feature extraction method is evaluated on four popular action datasets, namely KTH, HMDB51, UCF YouTube, and Hollywood2. Experimental results show that our method significantly outperforms other types of features, either hand-designed or machine-learned. © 2015 IEEE.
  Accession Number: 20150800556273
• Record 144 of
  
  Title:Optimal design of a space target acquisition optical system with small F-number
  
  Author(s):Yan, Peipei(1,2); She, Wenji(1); Liu, Kai(1); Duan, Jing(1); Jiang, Kai(1); Shan, Qiusha(1)
  Source: Proceedings of SPIE - The International Society for Optical Engineering  Volume: 10154  Issue:   DOI: 10.1117/12.2243586  Published: 2016  
  Abstract:A kind of space target acquisition optical system with small F-number was designed. The system had a working wavelength range of 0.45∼0.85μm, an effective focal length of 240 mm, a field of view is 2ω=3°, and an F-Number of F/2. The system characteristic is that the structure is simple. And the especial requirements of the spot, energy concentration, distortion and lateral color etc. are all satisfied. The primary and secondary mirrors are all spheres, so the difficulty and cost of machining are reduced. Moreover, the temperature characteristic of the system is analyzed. The temperature request is satisfied. © 2016 SPIE.
  Accession Number: 20170503310048