2021

• Record 13 of
  
  Title:Generic Demosaicking Method for Multispectral Filter Arrays Based on Adaptive Frequency Domain Filtering
  
  Author(s):Wang, Zecheng(1,2); Zhang, Geng(1); Hu, Bingliang(1)
  Source: ACM International Conference Proceeding Series  Volume:   Issue:   DOI: 10.1145/3502814.3502825  Published: October 15, 2021  
  Abstract:Multispectral filter arrays (MSFAs) are widely applied to achieve snapshot multispectral imaging on a single image sensor, which causes incomplete data of each channel in the original captured image, and thus a process of estimating missing data named "demosaicking"is needed for high spatial resolution imaging. In a multispectral imaging system equipped with MSFA, as the number of spectral channels increases, the lack of data in the original captured image becomes severer, which brings great challenges to the demosaicking process, and thus classical demosaicking methods for MSFAs often fail to satisfy both reconstructed image quality and computational efficiency. In this paper, we propose a generic demosaicking method for MSFAs based on adaptive frequency domain filtering (AFDF) which achieves high quality of reconstructed images with little computational cost. Experimental results demonstrate that our proposed demosaicking method outperforms the state-of-the-art methods in terms of both quality of reconstructed images and processing time. © 2021 ACM.
  Accession Number: 20221712036072
• Record 14 of
  
  Title:Low-velocity impact behaviors of glass fiber-reinforced polymer laminates embedded with shape memory alloy
  
  Author(s):Wang, Wenzhi(1,2); Zhao, Yueran(1,2); Chen, Shangjun(3); Jin, Xiaochao(4); Fan, Xueling(4); Lu, Chunsheng(5); Yang, Chengxing(6)
  Source: Composite Structures  Volume: 272  Issue:   DOI: 10.1016/j.compstruct.2021.114194  Published: September 15, 2021  
  Abstract:Shape memory alloy wires embedded glass fiber-reinforced polymer (SMA-GFRP) laminates have great potential in engineering applications. In this paper, low-velocity impact behaviors of SMA-GFRP laminates are investigated under different initial impact energies. Firstly, tensile tests are conducted on a single SMA wire and SMA-GFRP laminates to obtain their mechanical parameters. Then, finite element models are established to describe the mechanical behaviors of SMA-GFRP laminates. Finally, experiments and simulations are carried out to explore the low-velocity impact behaviors and damage mechanisms of SMA-GFRP laminates. The results show that, due to their excellent superelastic deformation and shape recovery ability, SMA wires can improve the damage tolerance and impact resistance of GFRP laminates. The damage patterns and mechanisms of SMA-GFRP laminates vary with the increase of initial impact energy. Under low and medium initial impact energies, deformation can be mostly recovered, while under high impact energy, laminates are almost penetrated and deformation cannot be recovered because of breakage of SMA wires. The damage area of laminates increases first and then decreases as the increase of impact energy. The findings provide a guidance for design and evaluation of SMA-GFRP laminates with low-velocity impact resistance. © 2021
  Accession Number: 20212410501599
• Record 15 of
  
  Title:Valuation Analysis of Chinese and American Listed Companies Based on Multiple Linear Regression and Grey Forecasting Model
  
  Author(s):Li, Jinke(1); Zhang, Qiang(2); Zhang, ZhiJun(3); Wang, Fan(2); Li, Xijie(3)
  Source: ACM International Conference Proceeding Series  Volume:   Issue:   DOI: 10.1145/3484274.3484308  Published: August 13, 2021  
  Abstract:With the development of China's economic globalization, the stock market has gradually demonstrated its important position in the development of China's market economy. First, this paper selects the average market-to-sales ratio as the valuation level, uses an evaluation model to calculate the valuation level of the Chinese A-share market and the US NASDAQ market in 2018, and calculates the valuation premium or discount level of these two markets. Secondly, we establish a multiple linear regression model to quantitatively analyze the relationship between the valuation indicators and fundamental indicators and liquidity indicators of China A-shares and the US NASDAQ market. Then, a grey forecast model is established to predict and analyze the fundamental indicators and liquidity indicators of the Chinese A-share market and the US NASDAQ market in 2019. According to the forecast results, the valuation indicators of these two markets in 2019 are calculated. The results found that the valuation level of my country's first batch of sci-tech innovation board companies fluctuates around 5 times, which is smaller than that of the United States, indicating that China's stock market has greater potential. © 2021 ACM.
  Accession Number: 20214911287705
• Record 16 of
  
  Title:Deep neural network oriented evolutionary parametric eye modeling
  
  Author(s):Zheng, Yang(1,2); Fu, Hong(3); Li, Ruimin(1,2); Hsung, Tai-Chiu(4); Song, Zongxi(1); Wen, Desheng(1)
  Source: Pattern Recognition  Volume: 113  Issue:   DOI: 10.1016/j.patcog.2020.107755  Published: May 2021  
  Abstract:Comprehensive and accurate eye modeling is crucial to a variety of applications, including human-computer interaction, assistive technologies, and medical diagnosis. However, most studies focus on the localization of one or two components of eyes, such as pupil or iris, lacking a comprehensive eye model. We propose to model an eye image by a set of parametric curves. The set of curves are plotted on an eye image to form a Contour-Eye image. A deep neural network is trained to evaluate the fitness of the Contour-Eye image. Then an evolutionary process is conducted to search the best fitting curve set, guided by the trained deep neural network. Finally, an accurate eye model with optimized parametric curves is obtained. For the algorithm evaluation, a finely annotated eye dataset denoted as FAED-50 is established by us, which contains 2,498 eye images from 50 subjects. The experimental results on the FAED-50 and the relabeled CASIA datasets and comparison with the state-of-the-art methods demonstrate the effectiveness and accuracy of the proposed parametric model. © 2020
  Accession Number: 20204809527185
• Record 17 of
  
  Title:Design of secondary lens focusing mechanism for multispectral camera
  
  Author(s):Rui, Wang(1,2); Xiuqin, Su(1,3); Yongming, Qiao(1); Tao, Lv(1); Xuan, Wang(1,2); Kaidi, Wang(1,2); Yuan, Tian(1,2)
  Source: Proceedings of SPIE - The International Society for Optical Engineering  Volume: 11885  Issue:   DOI: 10.1117/12.2602300  Published: 2021  
  Abstract:Due to the influence of mechanical environment, large range temperature change and atmospheric pressure, the space multispectral camera has a certain amount of defocus in the optical imaging system. In order to improve the imaging quality of the multispectral camera, if the traditional CAM focusing mechanism is adopted, it is difficult to meet the requirements of high precision focusing for fast response due to its disadvantages such as large volume, poor efficiency and low precision. Therefore, a new focusing mechanism is designed in this paper, which is composed of rhomboid amplifier large piezoelectric ceramic actuator, flexible hinge support structure and high-precision capacitance sensor. The mechanism drives the flexible support guide structure of the parallelogram by means of a rhomboid amplifier large PZT actuator with three points distributed uniformly and symmetrically at 120°, so that the lens base can move in a straight line along the Z direction. The high precision capacitance sensor is used as the feedback element to ensure the focusing accuracy of the mechanism reaches nanometer level. The test results show that the focusing range of this mechanism is ±21.91um, the focusing speed is 438um/s, the focusing precision is 50nm and the tilt error is 1". © 2021 SPIE
  Accession Number: 20212810616400
• Record 18 of
  
  Title:Bio-Inspired Representation Learning for Visual Attention Prediction
  
  Author(s):Yuan, Yuan(1); Ning, Hailong(2,3); Lu, Xiaoqiang(2)
  Source: IEEE Transactions on Cybernetics  Volume: 51  Issue: 7  DOI: 10.1109/TCYB.2019.2931735  Published: July 2021  
  Abstract:Visual attention prediction (VAP) is a significant and imperative issue in the field of computer vision. Most of the existing VAP methods are based on deep learning. However, they do not fully take advantage of the low-level contrast features while generating the visual attention map. In this article, a novel VAP method is proposed to generate the visual attention map via bio-inspired representation learning. The bio-inspired representation learning combines both low-level contrast and high-level semantic features simultaneously, which are developed by the fact that the human eye is sensitive to the patches with high contrast and objects with high semantics. The proposed method is composed of three main steps: 1) feature extraction; 2) bio-inspired representation learning; and 3) visual attention map generation. First, the high-level semantic feature is extracted from the refined VGG16, while the low-level contrast feature is extracted by the proposed contrast feature extraction block in a deep network. Second, during bio-inspired representation learning, both the extracted low-level contrast and high-level semantic features are combined by the designed densely connected block, which is proposed to concatenate various features scale by scale. Finally, the weighted-fusion layer is exploited to generate the ultimate visual attention map based on the obtained representations after bio-inspired representation learning. Extensive experiments are performed to demonstrate the effectiveness of the proposed method. © 2019 IEEE.
  Accession Number: 20213310758679
• Record 19 of
  
  Title:Hyperspectral image super-resolution with self-supervised spectral-spatial residual network
  
  Author(s):Chen, Wenjing(1,2); Zheng, Xiangtao(1); Lu, Xiaoqiang(1)
  Source: Remote Sensing  Volume: 13  Issue: 7  DOI: 10.3390/rs13071260  Published: April 1, 2021  
  Abstract:Recently, many convolutional networks have been built to fuse a low spatial resolution (LR) hyperspectral image (HSI) and a high spatial resolution (HR) multispectral image (MSI) to obtain HR HSIs. However, most deep learning-based methods are supervised methods, which require sufficient HR HSIs for supervised training. Collecting plenty of HR HSIs is laborious and time-consuming. In this paper, a self-supervised spectral-spatial residual network (SSRN) is proposed to alleviate dependence on a mass of HR HSIs. In SSRN, the fusion of HR MSIs and LR HSIs is considered a pixel-wise spectral mapping problem. Firstly, this paper assumes that the spectral mapping between HR MSIs and HR HSIs can be approximated by the spectral mapping between LR MSIs (derived from HR MSIs) and LR HSIs. Secondly, the spectral mapping between LR MSIs and LR HSIs is explored by SSRN. Finally, a self-supervised fine-tuning strategy is proposed to transfer the learned spectral mapping to generate HR HSIs. SSRN does not require HR HSIs as the supervised information in training. Simulated and real hyperspectral databases are utilized to verify the performance of SSRN. © 2021 by the authors. Licensee MDPI, Basel, Switzerland.
  Accession Number: 20211410165526
• Record 20 of
  
  Title:All-optical radio frequency spectrum analyzer with a 5 THz bandwidth based on CMOS-compatible high-index doped silica waveguides
  
  Author(s):Moss, David J.(8); Li, Yuhua(1,2); Kang, Zhe(3,4); Zhu, Kun(2); Ai, Shiqi(2); Wang, Xiang(5); Davidson, Roy R.(5); Wu, Yan(1); Morandotti, Roberto(6); Little, Brent E.(7); Chu, Sai Tak(2)
  Source: TechRxiv  Volume:   Issue:   DOI: 10.36227/techrxiv.13635380  Published: January 25, 2021  
  Abstract:We report an all-optical radio-frequency (RF) spectrum analyzer with a bandwidth greater than 5 terahertz (THz), based on a 50-cm long spiral waveguide in a CMOS-compatible high-index doped silica platform. By carefully mapping out the dispersion profile of the waveguides for different thicknesses, we identify the optimal design to achieve near zero dispersion in the C-band. To demonstrate the capability of the RF spectrum analyzer, we measure the optical output of a femtosecond fiber laser with an ultrafast optical RF spectrum in the terahertz regime. © 2021, CC BY.
  Accession Number: 20220140442
• Record 21 of
  
  Title:Sub-Pixel Scanning High-Resolution Panoramic 3D Imaging Based on a SPAD Array
  
  Author(s):Xue, Ruikai(1,2); Kang, Yan(1); Zhang, Tongyi(1,2); Li, Lifei(1); Zhao, Wei(1,2)
  Source: IEEE Photonics Journal  Volume: 13  Issue: 4  DOI: 10.1109/JPHOT.2021.3103817  Published: August 1, 2021  
  Abstract:3D imaging based on single-photon avalanche diode (SPAD) arrays is of interest in many applications. Limited by the small array size and low fill-factor of the available SPAD array, it is difficult for SPAD array-based lidars to achieve high-resolution and high-quality 3D images currently. Here, we propose a sub-pixel continuously scanning panoramic 3D imaging method. This method uses sub-pixel scanning to increase spatial resolution and employs accumulating photon counts at the different pixel of the array for the same target point to improve 3D image quality. We have established an experimental system based on a 32 × 32 SPAD array and a diffraction optical element (DOE) with 32 × 32 beamlet. By sub-pixel scanning for a target 3 m away and accumulating photon counts, we demonstrated that the spatial resolution has been improved from 3 mm to 0.33 mm and the 3D image quality has been improved evidently. © 2021 Institute of Electrical and Electronics Engineers Inc.. All rights reserved.
  Accession Number: 20231613903210
• Record 22 of
  
  Title:Extended field of view of light-sheet fluorescence microscopy by scanning multiple focus-shifted Gaussian beam arrays
  
  Author(s):Liu, Chao(1,2); Bai, Chen(1); Yu, Xianghua(1); Yan, Shaohui(1); Zhou, Yuan(1,2); Li, Xing(1,2); Min, Junwei(1); Yang, Yanlong(1); Dan, Dan(1); Yao, Baoli(1,2)
  Source: Optics Express  Volume: 29  Issue: 4  DOI: 10.1364/OE.418707  Published: February 15, 2021  
  Abstract:Light-sheet fluorescence microscopy (LSFM) facilitates high temporal-spatial resolution, low photobleaching and phototoxicity for long-term volumetric imaging. However, when a high axial resolution or optical sectioning capability is required, the field of view (FOV) is limited. Here, we propose to generate a large FOV of light-sheet by scanning multiple focus-shifted Gaussian beam arrays (MGBA) while keeping the high axial resolution. The positions of the beam waists of the multiple Gaussian beam arrays are shifted in both axial and lateral directions in an optimized arranged pattern, and then scanned along the direction perpendicular to the propagation axis to form an extended FOV of light-sheet. Complementary beam subtraction method is also adopted to further improve axial resolution. Compared with the single Gaussian light-sheet method, the proposed method extends the FOV from 12 µm to 200 µm while sustaining the axial resolution of 0.73 µm. Both numerical simulation and experiment on samples are performed to verify the effectiveness of the method. © 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
  Accession Number: 20210809951158
• Record 23 of
  
  Title:Photonic microwave and RF channelizers based on Kerr micro-combs
  
  Author(s):Tan, M.(1); Xu, X.(2); Wu, J.(1); Boes, A.(3); Nguyen, T.G.(3); Chu, S.T.(4); Little, B.E.(5); Morandotti, R.(6); Mitchell, A.(3); Moss, D.J.(1)
  Source: Proceedings of SPIE - The International Society for Optical Engineering  Volume: 11685  Issue:   DOI: 10.1117/12.2584015  Published: 2021  
  Abstract:We review recent work on broadband RF channelizers based on integrated optical frequency Kerr microcombs combined with passive micro-ring resonator filters, with microcombs having channel spacings of 200GHz and 49GHz. This approach to realizing RF channelizers offers reduced complexity, size, and potential cost for a wide range of applications to microwave signal detection. © 2021 SPIE.
  Accession Number: 20212610569967
• Record 24 of
  
  Title:Neuromorphic processing at 11 Tera-OPs with soliton crystal Kerr microcombs
  
  Author(s):Tan, M.(1); Xu, X.(2); Wu, J.(1); Boes, A.(3); Corcoran, B.(2); Nguyen, T.(3); Chu, S.T.(4); Little, B.E.(5); Morandotti, R.(6); Mitchell, A.(3); Moss, D.J.(1)
  Source: LEOS Summer Topical Meeting  Volume: 2021-July  Issue:   DOI: 10.1109/SUM48717.2021.9505771  Published: 2021  
  Abstract:We report a new approach to ONNs based on integrated Kerr micro-combs that is programmable, highly scalable and capable of reaching ultra-high speeds. We demonstrate a single neuron perceptron at 11.9 Giga-OPS at 8 bits per OP, or 95.2 Gbps. We then demonstrate a convolutional accelerator operating beyond 11 TeraOPs/s (TOPs). We test the perceptron on handwritten-digit recognition and cancer-cell detection - achieving over 90% and 85% accuracy, respectively. © 2021 IEEE.
  Accession Number: 20220711629870