检索结果-内蒙古大学图书馆

AMTrans:Auto-Correlation Multi-Head Attention Transformer for Infrared Spectral Deconvolution

清华大学学报自然科学版（英文版） 2025年第3期30卷 1329-1341页

作者： Lei Gao Liyuan Cui Shuwen Chen Lizhen Deng Xiaokang Wang Xiaohong Yan Hu Zhu College of Electronic and Optical Engineering & College of Flexible Electronics(Future Technology) Nanjing University of Posts and TelecommunicationsNanjing 210003China Jiangsu Province Key Lab on Image Processing and Image Communication Nanjing University of Posts and TelecommunicationsNanjing 210003China National Engineering Research Center of Communication and Network Technology Nanjing University of Posts and TelecommunicationsNanjing 210003China School of Computer Science and Technology Hainan UniversityHaikou 570228China

Infrared spectroscopy analysis has found widespread applications in various fields due to advancements in technology and industry *** improve the quality and reliability of infrared spectroscopy signals,deconvolution is a crucial preprocessing *** by the transformer model,we propose an Auto-correlation Multi-head attention Transformer(AMTrans)for infrared spectrum sequence *** auto-correlation attention model improves the scaled dot-product attention in the *** utilizes attention mechanism for feature extraction and implements attention computation using the auto-correlation *** auto-correlation attention model is used to exploit the inherent sequence nature of spectral data and to effectively recovery spectra by capturing auto-correlation patterns in the *** proposed model is trained using supervised learning and demonstrates promising results in infrared spectroscopic *** comparing the experiments with other deconvolution techniques,the experimental results show that the method has excellent deconvolution performance and can effectively recover the texture details of the infrared spectrum.

关键词： spectroscopy spectral deconvolution transformer auto-correlation mechanism

来源：评论

学校读者我要写书评

暂无评论

M-RAT: a Multi-grained Retrieval Augmentation Transformer for image Captioning 17th

M-RAT: a Multi-grained Retrieval Augmentation Transformer f...

引用

17th Asian Conference on Computer Vision, ACCV 2024

作者： Song, Jiayan Pan, Renjie Zhou, Jun Yang, Hua Institute of Image Communication and Network Engineering Shanghai Jiao Tong University Shanghai200240 China Shanghai Key Lab of Digital Media Processing and Transmission Shanghai200240 China

ISBN: (纸本)9789819609079

Current encoder-decoder methods for image captioning mai-nly consist of an object detection module (two-stage), or rely on big models with large-scale datasets to improve the effectiveness, which leads to increasing computation costs and cannot introduce new external knowledge. In this paper, we propose a novel end-to-end method Multi-grained Retrieval Augmentation Transformer (M-RAT) that innovatively fuses retrieved text derived from a changeable datastore with input visual feature through a Multi-modal Aligned Encoder, and introduce a specialized attention mechanism, Multi-MSA, to exploit both local and global interactions for delicate fine-grained details. Additionally, we enhance the decoder generation ability by employing low-level and high-level fused embeddings. Experiments demonstrate that M-RAT achieves comparable performance to state-of-the-art baselines with remarkable accuracy and details, as well as showing excellent domain adaptability for novel objects. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Decoding

来源：评论

学校读者我要写书评

暂无评论

Low-light image enhancement algorithm using a residual network with semantic information

引用

The Journal of China Universities of Posts and Telecommunications 2022年第2期29卷 52-62,84页

作者： Duan Lian Tang Guijin Jiangsu Key Laboratory of Image Processing and Image Communication Nanjing University of Posts and TelecommunicationsNanjing 210003China

Aiming to solve the poor performance of low illumination enhancement algorithms on uneven illumination images,a low-light image enhancement(LIME)algorithm based on a residual network was *** algorithm constructs a deep network that uses residual modules to extract image feature information and semantic modules to extract image semantic information from different ***,a composite loss function was also designed for the process of low illumination image enhancement,which dynamically evaluated the loss of an enhanced image from three factors of color,structure,and *** ensures that the model can correctly enhance the image features according to the image semantics,so that the enhancement results are more in line with the human visual *** results show that compared with the state-of-the-art algorithms,the semantic-driven residual low-light network(SRLLN)can effectively improve the quality of low illumination images,and achieve better subjective and objective evaluation indexes on different types of images.

关键词： image enhancement convolutional neural network(CNN) residual learning image semantics

来源：评论

学校读者我要写书评

暂无评论

Free Viewpoint Action Recognition based on Self-similarities

Free Viewpoint Action Recognition based on Self-similarities

引用

2012 IEEE 11th International Conference on Signal processing(ICSP 2012)

作者： Jiao Wang Changhong Chen Xiuchang Zhu Jiangsu Provincial Key Lab of Image Processing and Image Communication Nanjing University of Posts and Telecommunications

ISBN: (纸本)9781467321969

Action recognition is an important topic in computer vision and most current work focuses on view-dependent representations. In this paper, we develop a novel free viewpoint action recognition based on Self-similarity matrix (SSM), which tends to be stable across views. We choose Local Self-similarity (LSS) descriptor as our low-level feature, then SSM is calculated by computing the similarity between any pair of frame features. Each video sequence is represented using a diagonal descriptor vector extracted from the SSM. Support Vector Machines (SVM) is employed for classification. The encouraging experimental results on the public IXMAS multi-view data set demonstrate effectiveness of the proposed method.

关键词： local self-similarity descriptor self-similarity matrix diagonal feature action recognition view independent

来源：评论

学校读者我要写书评

暂无评论

Low-light image enhancement based on Retinex theory and dual-tree complex wavelet transform

引用

Optoelectronics Letters 2018年第6期14卷 470-475页

作者： YANG Mao-xiang TANG Gui-jin LIU Xiao-hua WANG Li-qian CUI Zi-guan LUO Su-huai Jiangsu Key Lab on Image Processing & Image Communication Nanjing University of Posts and Telecommunications School of Electrical Engineering and Computing University of NewcastleNSW 2308 Australia

In order to enhance the contrast of low-light images and reduce noise in them, we propose an image enhancement method based on Retinex theory and dual-tree complex wavelet transform(DT-CWT). The method first converts an image from the RGB color space to the HSV color space and decomposes the V-channel by dual-tree complex wavelet transform. Next, an improved local adaptive tone mapping method is applied to process the low frequency components of the image, and a soft threshold denoising algorithm is used to denoise the high frequency components of the image. Then, the V-channel is rebuilt and the contrast is adjusted using white balance method. Finally, the processed image is converted back into the RGB color space as the enhanced result. Experimental results show that the proposed method can effectively improve the performance in terms of contrast enhancement, noise reduction and color reproduction.

关键词： Retinex theory dual-tree complex wavelet transform image enhancement

来源：评论

学校读者我要写书评

暂无评论

Improved SAI method using non-local spatial constraint for image interpolation

Improved SAI method using non-local spatial constraint for i...

引用

2011 International Conference on Wireless communications and Signal processing, WCSP 2011

作者： Bian, Jueping Gan, Zongliang Zhang, Mengcheng Zhu, Xiuchang Jiangsu Province Key Lab on Image Processing and Image Communication NUPT Nanjing China

ISBN: (纸本)9781457710100

This paper proposes an improved image interpolation method based on the soft-decision adaptive interpolation (SAI) algorithm. Natural images often contain repeatable patterns and structures throughout the image, which is called non-local property. We can use this non-local strategy to improve the interpolation quality by better estimating the model parameters and Lagrangian multiplier. There are two steps in our method. In the first step, similar patches of the given block are found in the initialized high resolution image, and the model parameters can be determined properly using the expanded piecewise auto regression (PAR) model and non-local spatial constraint. In the second step, the self-similarity of patches across the high and low resolution images is exploited to solve the Lagrangian multiplier , thus to make the data estimation robust. Experiments indicate that the improved method can achieve good results both subjectively and objectively. © 2011 IEEE.

关键词： Interpolation

来源：评论

学校读者我要写书评

暂无评论

A Novel Multi-Stream Fusion Network for Underwater image Enhancement

引用

China communications 2024年第2期21卷 166-182页

作者： Guijin Tang Lian Duan Haitao Zhao Feng Liu Jiangsu Key Laboratory of Image Processing and Image Communication Nanjing University of Posts and TelecommunicationsNanjing 210003China Jiangsu Key Laboratory of Wireless Communications Nanjing University of Posts and TelecommunicationsNanjing 210003China

Due to the selective absorption of light and the existence of a large number of floating media in sea water, underwater images often suffer from color casts and detail blurs. It is therefore necessary to perform color correction and detail restoration. However,the existing enhancement algorithms cannot achieve the desired results. In order to solve the above problems, this paper proposes a multi-stream feature fusion network. First, an underwater image is preprocessed to obtain potential information from the illumination stream, color stream and structure stream by histogram equalization with contrast limitation, gamma correction and white balance, respectively. Next, these three streams and the original raw stream are sent to the residual blocks to extract the features. The features will be subsequently fused. It can enhance feature representation in underwater images. In the meantime, a composite loss function including three terms is used to ensure the quality of the enhanced image from the three aspects of color balance, structure preservation and image smoothness. Therefore, the enhanced image is more in line with human visual ***, the effectiveness of the proposed method is verified by comparison experiments with many stateof-the-art underwater image enhancement algorithms. Experimental results show that the proposed method provides superior results over them in terms of MSE,PSNR, SSIM, UIQM and UCIQE, and the enhanced images are more similar to their ground truth images.

关键词： image enhancement multi-stream fusion underwater image

来源：评论

学校读者我要写书评

暂无评论

Compressed Video Super-resolution Reconstruction Based on Regularized Algorithm

Compressed Video Super-resolution Reconstruction Based on Re...

引用

2006 8th International Conference on Signal processing

作者： Xu Zhong-qiang Gan Zongliang Zhu Xiu-chang Information Industry Ministry and Jiangsu Province Key Lab of Image Processing & Image Communication Nanjing University of Posts & Telecommunications

ISBN: (纸本)0780397371

Estimating high-resolution(HR) video from a sequence of low-resolution(LR) compressed observations is the focus of this *** on the theory of regularization,this paper proposes a new form of regularized cost function to control the within-channel balance between received data and prior information,and a channel weight coefficient to control the cross-channel *** LR frames are adaptively weighted according to their reliability and the regularization parameter is simultaneously estimated for each channel with ameliorating artifacts in compressed *** iterative gradient descent algorithm is utilized to reconstruction the HR *** results demonstrate that the proposed algorithm has an improvement in terms of both objective and subjective quality.

关键词： Regularization multichannel super-resolution reconstruction compressed video

来源：评论

学校读者我要写书评

暂无评论

Speaker conversion using kernel non-negative matrix factorization

引用

The Journal of China Universities of Posts and Telecommunications 2017年第5期24卷 60-67页

作者： Xu Qinyu Lu Guanming Yan Jingjie Li Haibo Cheng Xiao College of Telecommunications and Information Engineering Nanjing University of Posts and Telecommunications Nanjing 210003 China Jiangsu Province Key Laboratory on Image Processing and Image Communication Nanjing 210003 China

Voice conversion （VC） based on Gaussian mixture model （GMM） is the most classic and common method which converts the source spectrum to target spectrum. However this method is prone to over-fitting because of its frame-by-frame conversion. The VC with non-negative matrix factorization （NMF） is presented in this paper, which can keep spectrum from over-fitting by adjusting the size of basis vector （dictionary）. In order to realize the non-linear mapping better, kernel NMF （KNMF） is adopted to achieve spectrum mapping. In addition, to increase the accuracy of conversion, KNMF combined with GMM （GKNMF） is also introduced into VC. In the end, KNMF, GKNMF, GMM, principal component regression （PCR）, PCR combined with GMM （GPCR）, partial least square regression （PLSR）, NMF correlation-based frequency warping （NMF-CFW） and deep neural network （DNN） methods are compared with each other. The proposed GKNMF gets better performance in both objective evaluation and subjective evaluation.

关键词： VC kemel NMF spectrum mapping

来源：评论

学校读者我要写书评

暂无评论

image Demosaicing by Non-local Similarity and Local Correlation

Image Demosaicing by Non-local Similarity and Local Correlat...

引用

2012 IEEE 11th International Conference on Signal processing(ICSP 2012)

作者： Wang Guo-gang Zhu Xiu-chang Gan Zong-liang Province Key Lab on Image Processing and Image Communication Nanjing University of Posts and Telecommunications

ISBN: (纸本)9781467321969

Conventional single-chip digital cameras use color filter arrays(CFA) to sample different spectral components. image demosaicing is a problem of interpolating these data to complete red, green, and blue values for each image pixel, to produce an RGB image. Many color demosaicing(CDM) methods assume that the high local spatial redundancy exists among the color samples. Such an assumption, however, may be fail for images with high color saturation and sharp color transitions. This paper presents an adaptive demosaicing algorithm by exploiting both the non-local similarity and the local correlation(NLS-LC) in the color filter array image. First, the most flattest nonlocal image patches are searched in the searching window centered on the estimated pixel. Second, the patch, which is the most similar to the current patch, is selected among the most smoothest nonlocal patches. Third, according to the similar degree and the local correlation degree, the obtained nonlocal image patch and the current patch are adaptively chosen to estimate the missing color samples. Experimental results indicate that the proposed method exhibits superior performance over many state-of-the-art color interpolation methods.

关键词： image demosaicing, image interpolation, local correlation, non-local similarity

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：