2021
2021
-
Record 1 of
Title:Delving into rectifiers in style-based image translation
Author(s):Zhang, Yipeng(1,2,3); Hu, Bingliang(1,2); Ning, Hailong(4); Wang, Quang(1,2)Source: arXiv Volume: Issue: DOI: null Published: November 20, 2021Abstract:While modern image translation techniques can create photorealistic synthetic images, they have limited style controllability, thus could suffer from translation errors. In this work, we show that the activation function is one of the crucial components in controlling the direction of image synthesis. Specifically, we explicitly demonstrated that the slope parameters of the rectifier could change the data distribution and be used independently to control the direction of translation. To improve the style controllability, two simple but effective techniques are proposed, including Adaptive ReLU (AdaReLU) and structural adaptive function. The AdaReLU can dynamically adjust the slope parameters according to the target style and can be utilized to increase the controllability by combining with Adaptive Instance normalization (AdaIN). Meanwhile, the structural adaptative function enables rectifiers to manipulate the structure of feature maps more effectively. It is composed of the proposed structural convolution (StruConv), an efficient convolutional module that can choose the area to be activated based on the mean and variance specified by AdaIN. Extensive experiments show that the proposed techniques can greatly increase the network controllability and output diversity in style-based image translation tasks.MSC Codes 68T07 ? 2021, CC BY-NC-ND.Accession Number: 20210398474 -
Record 2 of
Title:Simulation and analysis of atmospheric scattering in stray light testing for point source transmittance
Author(s):Ma, Zhanpeng(1,2); Chen, Qinfang(1); Wang, Hu(1,2)Source: Applied Optics Volume: 60 Issue: 2 DOI: 10.1364/AO.409656 Published: January 10, 2021Abstract:We propose a full-band model to quantitatively calculate terrestrial atmospheric scattering in stray light testing based on the Monte Carlo algorithm. Measurements are conducted using two classifications of air cleanliness at off-axis angles from 35? to 90?. Corresponding simulations of Mie scattering and Rayleigh scattering are used for a comparison with the measured values. The relative root mean square deviation of the simulation from the measurement result is 3.72% and 24.1% for Mie scattering and Rayleigh scattering, respectively. This exhibits excellent agreement between the measured and predicted values for a 26? full-angle baffle when illuminated by a 550 mm diameter collimated beam. ? 2021 Optical Society of AmericaAccession Number: 20211310150442 -
Record 3 of
Title:The phase uncertainty from the fringe contrast of interferogram in Doppler asymmetric spatial heterodyne spectroscopy
Author(s):Sun, Chen(1,2); Feng, Yutao(2); Fu, Di(1,2); Wang, Pengchong(2); Sun, Jian(2); Bai, Qinglan(2)Source: Journal of Optics (United Kingdom) Volume: 23 Issue: 11 DOI: 10.1088/2040-8986/ac2462 Published: November 2021Abstract:The Doppler asymmetric spatial heterodyne spectroscopy is one of the primary techniques for measuring the upper atmospheric wind profile. In this work, the detailed derivation of the analytical expression of phase uncertainty was presented, including two significant parameters, fringe contrast and signal-to-noise ratio. The effectiveness of the re-parameterized analytical expression was proved using the numerical simulations and laboratory experiments, and both results are in good agreement with them from the analytical expression. Therefore, the re-parameterized analytical expression could be used to optimize the interferometer design and evaluating the instrument performance. ? 2021 The Author(s).Accession Number: 20214811229280 -
Record 4 of
Title:Star identification based on multilayer voting algorithm for star sensors
Author(s):Liu, Meiying(1,2); Wei, Xin(1,2); Wen, Desheng(1); Wang, Hu(1,2)Source: Sensors Volume: 21 Issue: 9 DOI: 10.3390/s21093084 Published: May 1, 2021Abstract:This paper describes the multilayer voting algorithm, a novel autonomous star identification method for spacecraft attitude determination. The proposed algorithm includes two processes: an initial match process and a verification process. In the initial match process, a triangle voting scheme is used to acquire candidates of the detected stars, in which the triangle unit is adopted as the basic voting unit. During the identification process, feature extraction is implemented, and each triangle unit is described by its singular values. Then the singular values are used to search for candidates of the imaged triangle units, which further improve the efficiency and robustness of the algorithm. After the initial match step, a verification method is applied to eliminate incorrect candidates from the initial results and then outputting the final match results of the imaged stars. Experiments show that our algorithm has more robustness to position noise, magnitude noise, and false stars than the other three algorithms, the identification speed of our algorithm is largely faster than the geometric voting algorithm and optimized grid algorithm. However, it takes more memory, and SVD also seems faster. ? 2021 by the authors. Licensee MDPI, Basel, Switzerland.Accession Number: 20211810282267 -
Record 5 of
Title:Spectral-Spatial Joint Sparse NMF for Hyperspectral Unmixing
Author(s):Dong, Le(1,3); Yuan, Yuan(2,4); Lu, Xiaoqiang(1)Source: IEEE Transactions on Geoscience and Remote Sensing Volume: 59 Issue: 3 DOI: 10.1109/TGRS.2020.3006109 Published: March 2021Abstract:The nonnegative matrix factorization (NMF) combining with spatial-spectral contextual information is an important technique for extracting endmembers and abundances of hyperspectral image (HSI). Most methods constrain unmixing by the local spatial position relationship of pixels or search spectral correlation globally by treating pixels as an independent point in HSI. Unfortunately, they ignore the complex distribution of substance and rich contextual information, which makes them effective in limited cases. In this article, we propose a novel unmixing method via two types of self-similarity to constrain sparse NMF. First, we explore the spatial similarity patch structure of data on the whole image to construct the spatial global self-similarity group between pixels. And according to the regional continuity of the feature distribution, the spectral local self-similarity group of pixels is created inside the superpixel. Then based on the sparse expression of the pixel in the subspace, we sparsely encode the pixels in the same spatial group and spectral group respectively. Finally, the abundance of pixels within each group is forced to be similar to constrain the NMF unmixing framework. Experiments on synthetic and real data fully demonstrate the superiority of our method over other existing methods. ? 1980-2012 IEEE.Accession Number: 20211010019659 -
Record 6 of
Title:Scientific objectives and payloads of Tianwen-1, China's first Mars exploration mission
Author(s):Zou, Yongliao(1,2); Zhu, Yan(1); Bai, Yunfei(1); Wang, Lianguo(1); Jia, Yingzhuo(1); Shen, Weihua(1); Fan, Yu(1); Liu, Yang(1,2); Wang, Chi(1,2); Zhang, Aibing(1); Yu, Guobin(3); Dong, Jihong(4); Shu, Rong(5); He, Zhiping(5); Zhang, Tielong(6); Du, Aimin(7); Fan, Mingyi(8); Yang, Jianfeng(9); Zhou, Bin(10); Wang, Yi(11); Peng, Yongqing(12)Source: Advances in Space Research Volume: 67 Issue: 2 DOI: 10.1016/j.asr.2020.11.005 Published: January 15, 2021Abstract:This paper describes the scientific objectives and payloads of Tianwen-1, China's first exploration mission to Mars. An orbiter, carrying a lander and a rover, lifted-off in July 2020 for a journey to Mars where it should arrive in February 2021. A suite of 13 scientific payloads, for in-situ and remote sensing, autonomously commanded by integrated payload controllers and mounted on the orbiter and the rover will study the magnetosphere and ionosphere of Mars and the relation with the solar wind, the atmosphere, surface and subsurface of the planet, looking at the topography, composition and structure and in particular for subsurface ice. The mission will also investigate Mars climate history. It is expected that Tianwen-1 will contribute significantly to advance our scientific knowledge of Mars. ? 2020 COSPARAccession Number: 20204909586443 -
Record 7 of
Title:TTH-RNN: Tensor-Train Hierarchical Recurrent Neural Network for Video Summarization
Author(s):Zhao, Bin(1); Li, Xuelong(1); Lu, Xiaoqiang(2)Source: IEEE Transactions on Industrial Electronics Volume: 68 Issue: 4 DOI: 10.1109/TIE.2020.2979573 Published: April 2021Abstract:Although a recurrent neural network (RNN) has achieved tremendous advances in video summarization, there are still some problems remaining to be addressed. In this article, we focus on two intractable problems when applying an RNN to video summarization: first the extremely large feature-to-hidden matrices. Since video features are usually in a high-dimensional space, it leads to extremely large feature-to-hidden mapping matrices in the RNN model, which increases the training difficulty. Second, the deficiency in long-range temporal dependence exploration. Most videos contain thousands of frames at least, which is such a long sequence that traditional RNNs cannot deal well with. Facing the abovementioned two problems, we develop a tensor-train hierarchical recurrent neural network (TTH-RNN) for the video summarization task. It contains a tensor-train embedding layer to avert the large feature-to-hidden matrices, together with a hierarchical structure of an RNN to explore the long-range temporal dependence among video frames. Practically, the experimental results on four benchmark datasets, including SumMe, TVsum, MED, and VTW, have demonstrated the excellent performance of a TTH-RNN in video summarization. ? 1982-2012 IEEE.Accession Number: 20205209675945 -
Record 8 of
Title:Expression Recognition Algorithm Based on Infrared Image
Author(s):Cui, Ying(1); Qiu, Shi(2)Source: Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST Volume: 388 Issue: DOI: 10.1007/978-3-030-82565-2_26 Published: 2021Abstract:It’s important to recognize facial expressions in social communication. To solve the problem that facial expression recognition by visible light is vulnerable to interference, we built a model from the perspective of thermal infrared. Based on the distribution characteristics of thermal infrared, the face region is firstly located by building a multi-projection model toward color. Then, the level set function of the local Gaussian fitting model was optimized, the regular term was removed, and the larger iteration step size was selected to achieve accurate face segmentation on the premise of segmentation accuracy. Finally, based on the structure of traditional deep learning network, the characteristics of DPN and CBAM network are given full play to realize expression recognition by thermal infrared images. ? 2021, ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering.Accession Number: 20213910959507 -
Record 9 of
Title:UAV-borne Biomimetic Curved Compound-eye Imaging System for Velocity Measurement
Author(s):Xu, Huangrong(1,2); Liu, Jinheng(1,2); Zhang, Yuanjie(1,2); Wu, Dengshan(1); Fan, Hao(1,2); Feng, Xiangpeng(1); Zhang, Geng(1,2); Hu, Bingliang(1,2); Yu, Weixing(1,2)Source: Guangzi Xuebao/Acta Photonica Sinica Volume: 50 Issue: 9 DOI: 10.3788/gzxb20215009.0911004 Published: September 25, 2021Abstract:Based on the visual advantages of the biomimetic compound eye, the biomimetic curved compound-eye is applied to the unmanned aerial vehicle-borne photoelectric detection system to realize the airborne wide field-of-view and high-resolution detection of moving objects purposes. According to the characteristics of the biological compound eye, a lens array arranged on the curved surface in a hexagon is designed as compound-eye lens, combined with an optical relay subsystem and a CMOS image sensor to form a biomimetic curved compound-eye imaging and velocity measurement system. The developed biomimetic curved compound-eye imaging system has a field of view of 98°×98°, a system focal length of 5 mm, an angular resolution of 1.8 mrad and an F-number of 3.5. The size of the system is Ф123 mm ×195 mm, and the weight is 1.35 kg. According to the imaging principle of the biomimetic compound-eye and by taking advantage of the overlapping field of views between adjacent ommatidia, the velocity measurement principle of the biomimetic curved compound-eye is proposed. The velocity measurement experiment of a moving car shows that the velocity measurement method can effectively improve the test reliability and accuracy of the moving target. ? 2021, Science Press. All right reserved.Accession Number: 20213910941912 -
Record 10 of
Title:An improved fusion method of infrared and visible images based on fusionGAN
Author(s):Yao, Zhiqiang(1,2); Guo, Huinan(2); Ren, Long(2)Source: Proceedings of SPIE - The International Society for Optical Engineering Volume: 11878 Issue: DOI: 10.1117/12.2599559 Published: 2021Abstract:Convolutional neural network is widely used in image fusion. However, the deep learning framework is only applied in some part of the fusion process in most existing methods. To generate a full end-to-end image fusion pipeline, a Y-shaped Generator model based on Generative Adversarial Network for infrared and visible image fusion is proposed. The idea of this method is to establish an adversarial game between the generator and the discriminator. The generator consisting of two Pyramid networks and three convolutional layers works as an autoencoder to improve the characteristic information of the fused images. As for the discriminator, it adopts a network structure similar to the Visual Geometry Group (VGG) network. The loss function uses the ratio loss to control the trade-off among generation loss and reconstruction loss. Results on publicly available datasets demonstrate that our method can improve the quality of detail information and sharpen the edge of infrared targets. ? 2021 SPIEAccession Number: 20212810622827 -
Record 11 of
Title:Design of telecentric relay optical system with broadband and real entrance pupil
Author(s):Chang, Lingying(1); Zhang, Qiang(1); Qiu, Yuehong(2); Zhang, Rong(1)Source: Hongwai yu Jiguang Gongcheng/Infrared and Laser Engineering Volume: 50 Issue: 10 DOI: 10.3788/IRLA20210091 Published: October 25, 2021Abstract:The relay optical system was widely used in optical systems such as infrared spectral imaging system, light field imaging system, optical microscopy system, polarization interference imaging system, compound eye imaging system, ring-belt panoramic optical system, multi-scale imaging system and head-mounted enhanced display system, etc, which can link up, match pupil and deflect optical systems. The structures of the existing relay optical system were studied. The design method of telecentric off-axis three-mirror optical system with front aperture and the description method of free-form surface were introduced. According to the design parameters, the telecentric relay optical system with broadband and real entrance pupil was completed. The system was an off-axis three-mirror optical system structure, and each mirror was a free-form surface described by XY polynomial. The results of CODEV software simulation show that the MTF of the system is close to the diffraction limit, the distortion is less than 1%, and the imaging quality is good with the working spectral range of 0.4-5.0 μm, f'=400 mm, F/3 and 2ω=8°. ? 2021, Editorial Board of Journal of Infrared and Laser Engineering. All right reserved.Accession Number: 20214311083836 -
Record 12 of
Title:High-accuracy twist measurement based on the spherical wave Talbot effect for a bi-grid modulation collimator
Author(s):Liu, Shengrun(1,2); Xue, Bin(1); Cheng, Ying(1); Yang, Jianfeng(1)Source: Applied Optics Volume: 60 Issue: 22 DOI: 10.1364/AO.429980 Published: August 1, 2021Abstract:The bi-grid modulation collimator is a significant way for imaging solar flares in hard x rays. It implements many subcollimators that consist of separated grid pairs (so-called front grid and rear grid) whose line orientations are parallel. However, when the twist of the front grid with respect to the other will be induced during testing of the bi-grid modulation collimator in the ground verification phase, the line orientation of the grid pairs are no longer parallel. Knowledge of the relative twist between the rear grid and the front grid is very helpful in improving the imaging quality of the bi-grid collimator. However, because of the wide spacing between grid pairs and the requirement of high measurement accuracy, it is a challenge to measure the relative twist. To meet this demand, a method based on the spherical wave Talbot effect is proposed. The Talbot images of the front grid and the rear grid are imaged on the same plane, respectively, through two proper spherical waves. The relative twist can be figured out through the angle between the stripes in the Talbot images of the front grid and the rear grid. In experiments, the measurement accuracy of the relative twist angle can reach 9 arcsec in the range of 370 arcsec. It demonstrates that this method can effectively measure the relative twist between the grid pairs with very high accuracy. ? 2021 Optical Society of AmericaAccession Number: 20213110716058