检索结果-内蒙古大学图书馆

Underwater image Enhancement Method Based on Dynamic Heterogeneous Feature Fusion neural Network

IEEE ACCESS 2022年 10卷 91816-91827页

作者： Zhou, Xuejun Zhang, Jiachen Zhou, Fangyuan Yanan Univ Coll Phys & Elect Informat Yanan 716000 Shanxi Peoples R China Calif Amer Univ Dept Comp Sci Los Angeles CA USA

In recent years, signal and image processing based on fractional calculus has attracted extensive attention. Aiming at the serious problem of gray-scale loss in the existing pseudo color methods in high gray-scale image enhancement, a pseudo color enhancement algorithm suitable for Dynamic heterogeneous feature fusion neural network is proposed, and the traditional jet, HSV and rainbow coding are improved. Firstly, bit depth quantization is performed on the high-level gray image;Secondly, color enhancement is realized by using the constructed high gray-scale enhancement algorithm;Then, combined with the convolution neural network, the compact learning method is used to extract the features of the multi-scale image, and the jump connection is used to prevent the gradient dispersion and overcome the fog blur effect of the underwater image The style cost function is used to learn the correlation between various channels of color image, improve the color correction ability of the model, and overcome the problem of color distortion of underwater image. Experimental results show that compared with traditional image enhancement methods, the proposed method has better comprehensive performance in subjective vision and objective indicators, and has advantages in dealing with underwater image enhancement. While improving the brightness of the image, the problem of color distortion and brightness blocking of the enhanced image is solved. The texture information of the image is effectively restored. The brightness distribution of the enhanced image can well restore the brightness distribution of the real shooting environment, which verifies that the algorithm has higher robustness.

关键词： Brightness image enhancement image color analysis Convolution Cameras Imaging Heuristic algorithms Deep learning Underwater tracking Deep learning image processing underwater image enhancement

来源：评论

学校读者我要写书评

暂无评论

Pseudo-Retinex decomposition-based unsupervised underwater image enhancement and beyond

引用

DIGITAL signal processing 2023年第1期137卷

作者： Shen, Zhen Xu, Haiyong Jiang, Gangyi Yu, Mei Du, Baozhen Luo, Ting Zhu, Zhongjie Ningbo Univ Sch Math & Stat Ningbo 315211 Peoples R China Ningbo Univ Fac Informat Sci & Engn Ningbo 315211 Peoples R China Zhejiang Wanli Univ Coll Informat & Intelligence Engn Ningbo 315211 Peoples R China

Underwater images suffer from color casts and low contrast degraded due to wavelength-dependent light scatter and abortion of the underwater environment. To effectively improve the quality of the underwater images, deep learning-based underwater image enhancement methods have been widely proposed. However, most deep learning-based underwater image enhancement methods rely heavily on paired datasets. Actually, obtaining distortion-free images as reference images is difficult in underwater imaging. To address this problem, a fully Unsupervised convolution neural network-based Underwater image Enhancement (UUIE) is proposed by pseudo-Retinex decomposition. The innovation of the proposed UUIE is to establish a relationship between the underwater imaging model and the Retinex model, then use terrestrial images to replace underwater images for training and estimate pseudo-illumination and pseudo-reflection maps through self-supervision using the pseudo-Retinex decomposition. The pseudo -reflection image and pseudo-illumination image are reconstructed by the pseudo-Retinex decomposition to obtain the enhanced image. Additionally, the proposed UUIE can also be extended to image dehazing and low-light enhancement with only one trained model. Experimental results on synthetic and real -world datasets demonstrate the effectiveness of the proposed UUIE quantitatively and qualitatively.(c) 2023 Elsevier Inc. All rights reserved.

关键词： Unsupervised underwater image enhancement Pseudo-Retinex decomposition Convolution neural network Low -light image enhancement image dehazing Retinex model

来源：评论

学校读者我要写书评

暂无评论

A gradient-free training approach for optical neural networks based on stochastic functions 13

A gradient-free training approach for optical neural network...

引用

Optoelectronic Devices and Integration XIII 2024

作者： Qi, Ji Wang, Shuang Liu, Zheng Zhang, Yunfan Cai, Xiaomin Liu, Tiegen Xu, Tianhua School of Precision Instruments and Opto-Electronics Engineering Tianjin University Tianjin300072 China Tianjin Key Laboratory of Space Environment Simulation Technology Tianjin300450 China School of Engineering University of Warwick CoventryCV4 7AL United Kingdom

ISBN: (纸本)9781510682009

Due to increasingly large computational resources, modern neural networks are severely constrained due to their processing speed and energy consumption. Optical neural networks (ONNs), which use photonic structures to process signals at the physical level as an alternative to the computation in the electronic domain provided by traditional neural networks, are an attractive approach to implementing ultra-high-speed, low-energy parallel computation. Nevertheless, current training processes for electronic domain neural networks are optimized from gradient-based training methods, such as backpropagation, not compatible with ONNs with gradient-free features. In this work, a stochastic function-based gradient-free training method, i.e., stochastic function direct feedback alignment (SF-DFA) is demonstrated and evaluated. SF-DFA trains a gradient-free system using stochastic matrices and functions to replace the weights and gradients of the nodes in neural networks. Thus, it is feasible to train ONNs without a prior knowledge of the photonic system and its gradients. In addition, implementing such training process on optical hardware is also known to be possible. A series of studies have been carried out for a spectral slicing neural network (SS-NN) architecture trained by SF-DFA. The SS-NN system uses bandpass filters embedded in optical fiber micro rings to enable slicing of the optical signal spectrum. Our results demonstrate that the training of ONN using SF-DFA can converge efficiently, with higher processing speed and lower energy consumption compared to back-propagation. © 2024 SPIE. All rights reserved.

关键词： Photonic devices

来源：评论

学校读者我要写书评

暂无评论

Detection of epileptic seizure using EEG signals analysis based on deep learning techniques

引用

CHAOS SOLITONS & FRACTALS 2024年 181卷

作者： Abdulwahhab, Ali H. Abdulaal, Alaa Hussein Al-Ghrairi, Assad H. Thary Mohammed, Ali Abdulwahhab Valizadeh, Morteza Altinbas Univ Dept Elect & Comp Engn Istanbul Turkiye Al Iraqia Univ Coll Engn Dept Elect Engn Baghdad Iraq Al Nahrain Univ Dept Comp Sci Baghdad Iraq Al Karkh Univ Sci Dept Remote Sensing Baghdad Iraq Urmia Univ Fac Elect & Comp Engn Dept Commun Engn Orumiyeh Iran

The brain neurons' electrical activities represented by Electroencephalogram (EEG) signals are the most common data for diagnosing Epilepsy seizure, which is considered a chronic nervous disorder that cannot be controlled medically using surgical operation or medications with more than 40 % of Epilepsy seizure case. With the progress and development of artificial intelligence and deep learning techniques, it becomes possible to detect these seizures over the observation of the non -stationary -dynamic EEG signals, which contain important information about the mental state of patients. This paper provides a concerted deep machine learning model consisting of two simultaneous techniques detecting the activity of epileptic seizures using EEG signals. The timefrequency image of EEG waves and EEG raw waves are used as input components for the convolution neural network (CNN) and recurrent neural network (RNN) with long- and short-term memory (LSTM). Two processing signal methods have been used, Short -Time Fourier Transform (STFT) and Continuous Wavelet Transformation (CWT), have been used for generating spectrogram and scalogram images with sizes of 77 x 75 and 32 x 32, respectively. The experimental results showed a detection accuracy of 99.57 %, 99.57 % using CWT Scalograms, and 99.26 %, 97.12 % using STFT spectrograms as CNN input for the Bonn University dataset and the CHB-MIT dataset, respectively. Thus, the proposed models provide the ability to detect epileptic seizures with high success compared to previous studies.

关键词： Epileptic seizure Deep learning Convolutional neural network Recurrent neural network Electroencephalogram EEG

来源：评论

学校读者我要写书评

暂无评论

Degradation Pattern of High Speed Roller Bearings Using a Data-Driven Deep Learning Approach

引用

JOURNAL OF signal processing SYSTEMS FOR signal image AND VIDEO TECHNOLOGY 2022年第12期94卷 1557-1568页

作者： Rathore, Maan Singh Harsha, S. P. Indian Inst Technol Roorkee Dept Mech & Ind Engn Adv Mech Vibrat Lab Roorkee 247667 Uttarakhand India

In this paper, a data-driven approach is utilized for bearing condition monitoring involving the classification of different operating states by processing the raw vibration data. The vibration responses are analyzed and preprocessed before input to 1D-RCNN (one-dimensional residual convolutional neural network). The comparison results are based on commonly implemented evaluation indices such as precision, recall, F1-score, and ROC plots. Hence, the results revealed the superiority of the proposed methodology and its efficacy in segregating the bearing lifetime data into different operating conditions. Furthermore, t-SNE (t-distributed stochastic neighbor embedding) technique is implemented to represent the precise discriminative learning ability of different layers of the network. The overall classification accuracy values are obtained as 97.2% for 1D-RCNN, 95.31% for 1D-CNN, 86.2%, 86.42%, and 87.4% for SVM, KNN, and DNN, respectively. Hence, the proposed methodology may be effectively implemented for bearing health monitoring utilizing deep learning networks as classifiers.

关键词： Condition monitoring Residual learning Receiver operating characteristics t-distributed stochastic neighbor embedding

来源：评论

学校读者我要写书评

暂无评论

DERnet: a deep neural network for end-to-end reconstruction in magnetic particle imaging

引用

PHYSICS IN MEDICINE AND BIOLOGY 2024年第1期69卷 015002-015002页

作者： Peng, Zhengyao Yin, Lin Sun, Zewen Liang, Qian Ma, Xiaopeng An, Yu Tian, Jie Du, Yang Inst Automat CAS Key Lab Mol Imaging Beijing Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence Beijing Peoples R China Beijing Key Lab Mol Imaging Beijing Peoples R China Shandong Univ Sch Control Sci & Engn Jinan Shandon Peoples R China Beihang Univ Key Lab Big Data Based Precis Med Minist Ind & Informat Technol Beijing Peoples R China Beihang Univ Sch Engn Med Beijing Peoples R China

Objective. Magnetic particle imaging (MPI) shows potential for contributing to biomedical research and clinical practice. However, MPI images are effectively affected by noise in the signal as its reconstruction is an ill-posed inverse problem. Thus, effective reconstruction method is required to reduce the impact of the noise while mapping signals to MPI images. Traditional methods rely on the hand-crafted data-consistency (DC) term and regularization term based on spatial priors to achieve noise-reducing and reconstruction. While these methods alleviate the ill-posedness and reduce noise effects, they may be difficult to fully capture spatial features. Approach. In this study, we propose a deep neural network for end-to-end reconstruction (DERnet) in MPI that emulates the DC term and regularization term using the feature mapping subnetwork and post-processing subnetwork, respectively, but in a data-driven manner. By doing so, DERnet can better capture signal and spatial features without relying on hand-crafted priors and strategies, thereby effectively reducing noise interference and achieving superior reconstruction quality. Main results. Our data-driven method outperforms the state-of-the-art algorithms with an improvement of 0.9-8.8 dB in terms of peak signal-to-noise ratio under various noise levels. The result demonstrates the advantages of our approach in suppressing noise interference. Furthermore, DERnet can be employed for measured data reconstruction with improved fidelity and reduced noise. In conclusion, our proposed method offers performance benefits in reducing noise interference and enhancing reconstruction quality by effectively capturing signal and spatial features. Significance. DERnet is a promising candidate method to improve MPI reconstruction performance and facilitate its more in-depth biomedical application.

关键词： magnetic particle imaging end-to-end reconstruction deep learning image reconstruction

来源：评论

学校读者我要写书评

暂无评论

DanceCam: atmospheric turbulence mitigation in wide-field astronomical images with short-exposure video streams

引用

MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETY 2024年第1期531卷 403-421页

作者： Bialek, Spencer Bertin, Emmanuel Fabbro, Sebastien Bouy, Herve Rivet, Jean-Pierre Lai, Olivier Cuillandre, Jean-Charles Univ Victoria Dept Phys & Astron Victoria BC V8W 3P2 Canada Canada France Hawaii Telescope Kamuela HI 96743 USA Univ Paris Cite Univ Paris Saclay CNRS CEAAIM F-91191 Gif sur Yvette France Natl Res Council Herzberg Astron & Astrophys Victoria BC V9E 2E7 Canada CNRS Lab Astrophys Bordeaux Allee Geoffroy St Hilaire F-33165 Pessac France Inst Univ France Paris France Univ Cote Azur CNRS Observ Cote Azur Lab JL Lagrange F-06304 Nice 4 France

We introduce a novel technique to mitigate the adverse effects of atmospheric turbulence on astronomical imaging. Utilizing a video-to-image neural network trained on simulated data, our method processes a sliding sequence of short-exposure (similar to 0.2 s) stellar field images to reconstruct an image devoid of both turbulence and noise. We demonstrate the method with simulated and observed stellar fields, and show that the brief exposure sequence allows the network to accurately associate speckles to their originating stars and effectively disentangle light from adjacent sources across a range of seeing conditions, all while preserving flux to a lower signal-to-noise ratio than an average stack. This approach results in a marked improvement in angular resolution without compromising the astrometric stability of the final image.

关键词： methods: data analysis methods: observational techniques: image processing software: simulations turbulence atmospheric effects

来源：评论

学校读者我要写书评

暂无评论

Not All Patches Are Equal: Hierarchical Dataset Condensation for Single image Super-Resolution

引用

IEEE signal processing LETTERS 2023年 30卷 1752-1756页

作者： Ding, Qingtang Liang, Zhengyu Wang, Longguang Wang, Yingqian Yang, Jungang Natl Univ Def Technol Coll Elect Sci & Technol Changsha 410073 Peoples R China Aviat Univ Air Force Changchun 130012 Peoples R China

Although the performance of single image super-resolution (SR) has been significantly improved with deep neural networks, existing methods commonly require millions of iterations for training, which not only limits their training efficiency, but also causes considerable energy consumption. In this paper, we comprehensively study the redundancy of existing training datasets and reveal that not all patches are equal for SR network training. We observe that a large percentage of patches with low textures or similar textures lead to high computation costs but make low contributions to SR performance. Then, we propose a dataset condensation method to remove these redundant patches hierarchically. Extensive experiments demonstrate that our dataset condensation method can effectively reduce the redundancy of SR datasets with a 90% condensation rate on DIV2K. With our condensed dataset, baseline networks can achieve significant improvement in terms of training efficiency while maintaining competitive accuracy.

关键词： image super-resolution (SR) dataset condensation training efficiency

来源：评论

学校读者我要写书评

暂无评论

Blur-CAM: Visual interpretation of deep convolutional networks based on Gaussian fuzzy

Blur-CAM: Visual interpretation of deep convolutional networ...

引用

2024 International Conference on image, signal processing, and Pattern Recognition, ISPP 2024

作者： Liu, Haoxiang School of Information Science Yunnan University Kunming650091 China

ISBN: (纸本)9781510680425

As deep learning continues to advance, the complexity of its network structures has also *** application of visualization techniques to elucidate the learning mechanisms of deep convolutional neural networks constitutes a crucial aspect of research in interpretable artificial intelligence and the field of deep *** visualization methods have emerged to interpret these black-box architectures, enabling analysis and comprehension of internal decision-making *** this paper, we introduce Blur-CAM, a method that employs Gaussian blur to process both masked and unmasked regions, enhancing object understanding and providing smoother optimization transitions for object localization in *** evaluate our approach on the ILSVRC 2012 and PASCAL VOC 2012 datasets, achieving commendable results. © 2024 SPIE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Spatiotemporal Backpropagation based on Channel Reward for Training High-Precision Spiking neural Network

引用

signal image AND VIDEO processing 2023年第7期17卷 3467-3476页

作者： Niu, Li-Ye Wei, Ying Liu, Yue Long, Jun-Yu Liu, Wen-Bo Northeastern Univ Coll Informat Sci & Engn Shenyang 110004 Peoples R China Peking Univ Informat Technol R&D Innovat Ctr Shaoxing Peoples R China

Spiking neural network (SNN) has attracted much attention due to its spatial-temporal information processing ability and high biological reliability. SNN is a hardware-friendly model based on event-driven and sparse triggering, which can achieve low power consumption and efficient information processing on neural morphological chips. However, SNN processes discrete spike trains, and its unique working mode makes it more difficult to train than traditional networks. Based on this, this paper proposes a spatiotemporal backpropagation (STBP) algorithm for directly training high-performance SNN. By narrowing the coding time window, we convert the leaky integrate-and-fire (LIF) into its explicit iterative version for direct SNN training. To solve the problem of non-differentiability of SNN discrete spikes, we introduce an approximate derivative of spike firing for gradient descent training. We propose a biologically reasonable channel reward (CR) mechanism integrated into the STBP algorithm. This method makes full use of the spatial domain (SD) and time domain (TD) information of input and does not require any additional complex techniques. We evaluate the proposed algorithm on traditional static MNIST, Fashion-MNIST datasets, and neuromorphological N-MNIST, DVS128 Gesture datasets. The experimental results show that the accuracy of this method in static and neuromorphological datasets exceeds the current advanced methods, and the time window used is small. This work provides a new perspective for directly training high-performance SNN.

关键词： Spiking neural network Spatiotemporal backpropagation Channel reward Directly training Biologically reasonable

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：