检索结果-内蒙古大学图书馆

2023 International Conference on Image, Signal processing, and Pattern Recognition, ISPP 2023

作者： Li, Quanyin Xiang, Rui Huang, Yan Wang, Guoyou National Key Laboratory of Science and Technology on Multispectral Information Processing School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

ISBN: (纸本)9781510666351

In recent years, deep learning-based methods have made a big splash in the field of single image super-resolution (SISR), and a lot of excellent research works on lightweight image super-resolution networks have emerged. This series of approaches point a way forward for lightweight SR network design, but they generally suffer from two problems: 1) slow convergence of the network 2) poor image reconstruction quality. In this paper, we firstly propose a simple and effective learning rate decay scheme without additional cost, and the new strategy can accelerate the convergence speed of the network, and the model obtained in the same training time is about 0.05 dB higher in PSNR test results. Secondly, we think about the performance bottleneck of RFDN and propose an enhanced full-channel attention (EFA) module with a mixture of channel and spatial attention, which can improve about 0.085 dB in PSNR compared to the CCA attention module. Finally, we propose a new feature distillation network based on the EFA module, called EFAFDN. With only 523k parameters, the EFAFDN(x2) is lighter than the RFDN with 534k parameters and has better SR performance than the RFDN-L with 626k parameters. Extensive experimental results show that EFAFDN performs slightly better than RLFN, the winning model of the main track of the NTIRE 2022 Efficient Super-Resolution Challenge. © 2023 SPIE.

关键词： Distillation

来源：评论

学校读者我要写书评

暂无评论

Ascent Guidance for Airbreathing Hypersonic Vehicle Based on Deep Neural Network and Pseudo-spectral Method 1

引用

Proceedings of the 4th International Conference on Neural Computing for Advanced Applications, NCAA 2023

作者： Fu, Wenzhe Wang, Bo Liu, Lei Wang, Yongji National Key Laboratory of Science and Technology on Multispectral Information Processing School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

ISBN: (数字)9789819958443

ISBN: (纸本)9789819958436

The hypersonic vehicle has received wide attention because of its high speed and large flight domain. However, the ascent guidance is quite challenging due to its high speed, strong nonlinear dynamics, uncertain parameters, and coupling between the thrust and aerodynamic forces. The traditional model-based guidance laws may have large deviations under uncertain parameters when the model is inaccurate. This paper proposes a data-driven guidance method that consists of deep neural networks and a pseudo-spectral optimization method. The trajectory generation under different uncertain parameters is formulated as an optimal control problem which is solved by the pseudo-spectral method offline. The trajectory database is established based on the vast optimal trajectories and used to train the deep neural network (DNN). Afterward, the ascent guidance law is given by the well-trained DNNs online to lessen the effect of the uncertain parameters and nonlinear characteristics. Meanwhile, a sequential calling strategy is put forward to invoke the multiple DNNs, in which the remaining flight time is designed as the switching variable. The aim is to let the latter networks correct the integral deviation generated by the approximation error of the former networks, and the high-precision terminal states can be achieved. The Mento-Carlo simulations are carried out to give robustness. The results show that the proposed method has the advantage of small terminal deviations and no parameter-tunning. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Hypersonic vehicles

来源：评论

学校读者我要写书评

暂无评论

Mid-long-wave infrared broadband absorber based on multi-layer plasma cavities metasurface 12

Mid-long-wave infrared broadband absorber based on multi-lay...

引用

SPIE 12th International Symposium on multispectral Image processing and Pattern Recognition, MIPPR 2023

作者： Duan, Shenghua Yang, Guang Zhang, Xinyu National Key Laboratory of Science & Technology on Multispectral Information Processing Huazhong University of Science & Technology Wuhan430074 China School of Artificial Intelligence & Automation Huazhong University of Science & Technology Wuhan430074 China

ISBN: (数字)9781510674929

ISBN: (纸本)9781510674912

Efficient absorption in the mid-long-wave infrared region is crucial for various applications, including thermal imaging and sensing, infrared spectroscopy, and camouflage. A multilayered metal-insulator-metal-based infrared plasmonic metamaterial absorber utilizing a stacked nanocavity meta-surface consisting of a double-layer cylindrical metal-insulator-metal structure (DCMIM) is proposed. This absorber demonstrates wideband absorption capabilities with an average absorption rate exceeding 75% in the 3-14μm range. To understand the spectral absorption mechanisms, a comprehensive analysis is conducted using the Finite Difference Time Domain and Fabry-Perot resonance model. The analysis reveals physical mechanisms such as strong coupling between plasmon resonances and phonon vibrations loss from the dielectric spacer, as well as localized surface plasmon resonance. It is worth noting that the quantities of plasma cavities significantly impacts the performance of the absorber. By adjusting the material type and thickness of the film layer, it becomes possible to effectively compensate for the performance limitations of a single dielectric layer material within specific wavelength bands. Experimental results, combined with theoretical analysis of the DCMIM structure, establish a new approach to the design of broadband infrared absorbers. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Metal insulator boundaries

来源：评论

学校读者我要写书评

暂无评论

Multi-scale reconstruction generative networks for thermal to visible image transformation

Multi-scale reconstruction generative networks for thermal t...

引用

2023 International Conference on Image, Signal processing, and Pattern Recognition, ISPP 2023

作者： Xiang, Rui Li, Quanyin Huang, Yan Wang, Guoyou National Key Laboratory of Science and Technology on Multispectral Information Processing School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

ISBN: (纸本)9781510666351

In recent years, the field of thermal to visible image transformation has made great progress with Generative Adversarial Network(GAN). Since GANs own the ability to fit the signal frequency, these methods are limited to using more GANs or designing sub-tasks to improve the quality of the generated image. Due to the large difference in the information frequency of individual objects in the image, GANs pay less attention to low-frequency information, which leads to edge distortion of dynamic objects such as people and vehicles in the generated image. Edge distortion has become an important issue for practical applications We propose a new structure and two new loss functions to complete the task of thermal to visible image transformation. Our network takes one thermal image as input and generates a visible RGB image as output. First, we down-sample inputs for three times to get multi-scale inputs. Then we build outputs in multi-scale form small size to large size and fuse the multi-scale outputs. In train stage, our two new losses better direct the network to converge. In order to prove the effectiveness of the proposed thermal to visible image transformation algorithm, we use the KASIT-MPD dataset, which includes real paired thermal and visible images. © 2023 SPIE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Tuning of near-field optical properties based on magneto-tip array super-surfaces 12

Tuning of near-field optical properties based on magneto-tip...

引用

SPIE 12th International Symposium on multispectral Image processing and Pattern Recognition, MIPPR 2023

作者： Chen, Cheng Gao, Zecheng Duan, Shenghua Yang, Guang Zhang, Chunyu Wang, Zhe Zhang, Xinyu National Key Laboratory of Science & Technology on Multispectral Information Processing Huazhong University of Science & Technology Wuhan430074 China School of Artificial Intelligence & Automation Huazhong University of Science & Technology Wuhan430074 China

ISBN: (数字)9781510674929

ISBN: (纸本)9781510674912

A method of light wave modulation on a magnetic nano-tip hypersurface was investigated within the wavelength range of 3-14 μm, and a hypersurface was constructed on a silicon wafer using magnetic nano-film to perform light control using the magnetization of the magnetic nano-tip to achieve the desired light absorption. Through the magnetic nanotip hypersurface, the incident light waves hypersurface can be highly confined in the arrayed magnetic nanotip to achieve near-field light enhancement and light focusing, and significantly reduce the infrared reflectivity to achieve efficient absorption of infrared light waves. The infrared reflectivity of the structured surface shows a tunable variation by adjusting the structural parameters of the magnetic nano-tip super-surface. The experiments show that the silicon-based magnetic film nano-tip array achieves stronger nanofocusing and higher absorption rate than the conventional nano-tip array based on metal film. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Surface plasmons

来源：评论

学校读者我要写书评

暂无评论

Application of nanotips metasurface in extending image depth of field 12

Application of nanotips metasurface in extending image depth...

引用

SPIE 12th International Symposium on multispectral Image processing and Pattern Recognition, MIPPR 2023

作者： Liu, Taige Duan, Shenghua Ji, Wuyang Zhang, Xinyu National Key Laboratory of Science & Technology on Multispectral Information Processing Huazhong University of Science & Technology Wuhan430074 China School of Artificial Intelligence & Automation Huazhong University of Science & Technology Wuhan430074 China

ISBN: (数字)9781510674929

ISBN: (纸本)9781510674912

Metasurfaces, enabling effective optical modulation of incident lightwave, can be employed as a promising candidate to implement different functions, thus significantly improving the lightwave controllability, validity and system integration. In this paper, the phase and intensity modulation, firstly, of the incident lightwave based on a single nanotip was researched. Then, the metasurface consist of nanotips with different structural parameters was designed according to the grayscale image obtained via diffractive network encoding, so as to realize the extending image depth of field. We demonstrate that the metasurface based on the nanotips antenna can achieve a large image depth range of field exceeding 50μm. Moreover, the maximum focal length can reach ∼200μm after passing through the arrayed nanotips. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Light modulation

来源：评论

学校读者我要写书评

暂无评论

Linear Refinement Attention for Road Damage Detection

Linear Refinement Attention for Road Damage Detection

引用

International Conference on Wireless Communications, Networking and Applications (WCNA 2022)

作者： Wang, Faquan Han, Shoudong National Key Laboratory of Science and Technology on Multispectral Information Processing School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

ISBN: (纸本)9789819939503

Object detection plays a vital role in computer vision. As a typical application of object detection, road damage detection has strong practicability. Road damage includes longitudinal crack, transverse crack, etc. Compared with universal objects, the length-width ratio of road damage is usually extreme, and the road damage appears like a line. Thus, in the paper, we utilize this characteristic of road damage to propose linear refinement attention. Our linear refinement attention consists of two parts: linear pooling and linear convolution. They are connected in series and refine the intermediate feature maps’ global and local features, respectively. They both first capture high-level semantic features, then transform the captured features into weight, and finally utilize the weight to refine the input. The most common dataset for road damage detection is Road Damage Dataset 2020 (RDD2020). According to our observation, many images in RDD2020 lack labels, leading to unreasonable evaluation. Thus, we optimized the dataset to ensure a professional and accurate evaluation. We conducted many experiments, and the results show that our method can capture the features of road damage more accurately and improve the mAP steadily compared with the baseline. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

3D hand pose and shape estimation from monocular RGB via efficient 2D cues

引用

Computational Visual Media 2024年第1期10卷 79-96页

作者： Fenghao Zhang Lin Zhao Shengling Li Wanjuan Su Liman Liu Wenbing Tao Hubei Key Laboratory of Medical Information Analysis and Tumor Diagnosis&Treatment School of Biomedical EngineeringSouth Central Minzu UniversityWuhan430074China National Key Laboratory of Science and Technology of Multi-spectral Information Processing School of Artificial Intelligence and AutomationHuazhong University of Science and TechnologyWuhan430074China

Estimating 3D hand shape from a single-view RGB image is important for many ***,the diversity of hand shapes and postures,depth ambiguity,and occlusion may result in pose errors and noisy hand *** full use of 2D cues such as 2D pose can effectively improve the quality of 3D human hand shape *** this paper,we use 2D joint heatmaps to obtain spatial details for robust pose *** also introduce a depth-independent 2D mesh to avoid depth ambiguity in mesh regression for efficient hand-image *** method has four cascaded stages:2D cue extraction,pose feature encoding,initial reconstruction,and reconstruction ***,we first encode the image to determine semantic features during 2D cue extraction;this is also used to predict hand joints and for ***,during the pose feature encoding stage,we use a hand joints encoder to learn spatial information from the joint ***,a coarse 3D hand mesh and 2D mesh are obtained in the initial reconstruction step;a mesh squeeze-and-excitation block is used to fuse different hand features to enhance perception of 3D hand ***,a global mesh refinement stage learns non-local relations between vertices of the hand mesh from the predicted 2D mesh,to predict an offset hand mesh to fine-tune the reconstruction *** and qualitative results on the FreiHAND benchmark dataset demonstrate that our approach achieves state-of-the-art performance.

关键词： hand 3D reconstruction deep learning image features 3D mesh

来源：评论

学校读者我要写书评

暂无评论

Toroidal composite liquid crystal microlens array co-driven by four independent signal voltages 12

Toroidal composite liquid crystal microlens array co-driven ...

引用

SPIE 12th International Symposium on multispectral Image processing and Pattern Recognition, MIPPR 2023

作者： Wang, Zhe Yang, Guang Zhang, Xinyu National Key Laboratory of Science & Technology on Multispectral Information Processing Huazhong University of Science & Technology Wuhan430074 China School of Artificial Intelligence & Automation Huazhong University of Science & Technology Wuhan430074 China

ISBN: (数字)9781510674929

ISBN: (纸本)9781510674912

A new type of liquid crystal microlens array (LCMLA) constructed by a single-layered LC materials is proposed. By controlling the applied signal voltage, a gradient electric field distribution is formed in the liquid crystal cavity, and the pointing vectors of the liquid crystal molecules are reoriented according to the formed electric field distribution, thus achieving the regulation of the light beam. The functional area of the basic annular integrated liquid crystal microlens is divided into three layers 1/4 The first layer is the main structural area, including four concentric ring electrode arrays with different radii, and the transparent ITO is selected as the electrode material;The middle layer is a silica insulation layer, and the main function is to insulate the first and third layers;The third layer is an aluminum metal electrode layer, which main function is to effectively isolate the electric field generated by the structural area of the first layer. Compared with traditional LC microlenses driven electrically 1/4 this structure mainly verifies the changes in the spatial electric field in the liquid crystal cavity under the action of the metal electrode baffle. The four concentric rings are modulated by independently driven-out signal voltages, and each ring produces a differentiated effect by applying different signal voltages. This research has laid a solid foundation for continuously developing LCMLA technology. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Microlenses

来源：评论

学校读者我要写书评

暂无评论

Study on the polarization response of aluminum gratings with graphene 12

Study on the polarization response of aluminum gratings with...

引用

SPIE 12th International Symposium on multispectral Image processing and Pattern Recognition, MIPPR 2023

作者： Zhu, Wenjie Zhang, Xinyu National Key Laboratory of Science & Technology on Multispectral Information Processing Huazhong University of Science & Technology Wuhan430074 China School of Artificial Intelligence & Automation Huazhong University of Science & Technology Wuhan430074 China

ISBN: (数字)9781510674929

ISBN: (纸本)9781510674912

When the wave-vector matching criteria is satisfied, the incident light can excite the surface plasmon on the surface of the metallic micro-nano arrays, as to form the resonant surface electromagnetic field. From the perspective of surface charge rearrangement, we propose interfinger-type metal grating and cover the grating with graphene to manipulate the polarization response of light waves. By using FDTD software, we simulate the transmittance of the structure to the infrared light with different polarization angles, and analyze the local electric field distribution at the edge of the grating. According to the simulation results, the metal grating with graphene can manipulate the light wave transmission well, absorbing the polarization component parallel to the grating's long axis and passing through the polarization component perpendicular to the grating's long axis. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Graphene

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：