31,589 Hits in 3.3 sec

The Emergence of Spectral Universality in Deep Networks [article]

Jeffrey Pennington, Samuel S. Schoenholz, Surya Ganguli
2018 arXiv   pre-print
For a variety of nonlinearities, our work reveals the emergence of new universal limiting spectral distributions that remain concentrated around one even as the depth goes to infinity.  ...  Recent work has shown that tight concentration of the entire spectrum of singular values of a deep network's input-output Jacobian around one at initialization can speed up learning by orders of magnitude  ...  The Emergence of Spectral Universality in Deep Networks: Supplementary Material 1 Review of free probability For what follows, we define the key objects of free probability.  ... 
arXiv:1802.09979v1 fatcat:cho6fd6mdfddfenrdfsv3m2iru

Deep Learning Applications for Hyperspectral Imaging: A Systematic Review

Akin Ozdemir, School of Natural Sciences, Department of Electrical and Electronics Engineering Bolu Abant Izzet Baysal University, 14280, Bolu, Turkey, Kemal Polat, Department of Electrical and Electronics Engineering, Faculty of Engineering, Bolu Abant Izzet Baysal University, 14280, Bolu, Turkey
2020 Journal of the Institute of Electronics and Computer  
Neural networks are also used as an unsupervised learning method. However, Deep Learning, a specialized method of artificial neural networks, is highly preferred due to its unique structure.  ...  A hyperspectral image consists of reflections in hundreds of different bands of the electromagnetic spectrum. Each object exhibits a unique reflection characteristic.  ...  In this study, especially deep learning methods will be discussed. 4 different models that emerged within the scope of deep learning were examined.  ... 
doi:10.33969/jiec.2020.21004 fatcat:a5snmltpefgijj57xplvofs2x4

Special focus on deep learning in remote sensing image processing

Feng Xu, Cheng Hu, Jun Li, Antonio Plaza, Mihai Datcu
2020 Science China Information Sciences  
Special focus on deep learning in remote sensing image processing * As a newly emerging technology, deep learning is a very promising field in big data applications.  ...  In the contribution "Deep-learning-based extraction of the animal migration patterns from weather radar images", Cui et al. use convolutional neural networks to classify and segment weather radar images  ... 
doi:10.1007/s11432-020-2810-x fatcat:pjdzl7a7wfh7bpizagqaydlyea

Spatial-Spectral Random Patches Network for Classification of Hyperspectral Images

Behnam Beirami, Mehdi Mokhtarzade
2019 Traitement du signal  
However, this network only relies on spectral bands in feature extraction, failing to make use of the information-rich spatial features.  ...  The results prove the superiority of the proposed method in the classification of hyperspectral images over some recent shallow and deep spatial-spectral classification techniques.  ...  As another issue, in many deep learning models, the deepest level of features (features from the last network layer) in the network is used in the Softmax layer for the analysis of HSI.  ... 
doi:10.18280/ts.360504 fatcat:ucd44ntflrdefo3cf3i4dfhlpy

3D-ANAS v2: Grafting Transformer Module on Automatically Designed ConvNet for Hyperspectral Image Classification [article]

Xizhe Xue, Haokui Zhang, Zongwen Bai, Ying Li
2021 arXiv   pre-print
Benefiting from the development of deep learning technologies, deep learning based HSI classification methods have achieved promising performance.  ...  In addition, to further improve the classification accuracy, we attempt to graft the emerging transformer module on the automatically designed ConvNet to adding global information to local region focused  ...  Due to the unique imaging characteristic of HSI, different objects in HSI may have similar spectral features and the same objects located in different locations may emerge with different spectral features  ... 
arXiv:2110.11084v1 fatcat:z3yij5i3mjcihcemzto4nw2giy

Detailed Technical Programme Schedule

2020 2020 Sixth International Conference on Parallel, Distributed and Grid Computing (PDGC)  
Chair(s): Dr.Himanshu Jindal, Jaypee University of Information Technology, Waknaghat 11.1: Emerging Applications of IoT, Deep learning and its impact on Image Dr.Rajinder Sandhu, Jaypee University of  ...  Neural Network Approach for The Gitanjali Wadhwa 08.2: Neural Networks and Deep Learning based Methods for Digital Image Raman Dugyala, Jawaharlal Nehru Technological University, Hyderabad Information  ... 
doi:10.1109/pdgc50313.2020.9315322 fatcat:4ndwytytovb7xkntz7bmaurj6a

Hyperspectral Images Classification Based on Multi-scale Residual Network [article]

Xiangdong Zhang, Tengjun Wang, Yun Yang
2020 arXiv   pre-print
The experimental results show that the overall classification accuracy of this method is 99.07% and 99.96% respectively in the data set of Indian Pines and Pavia University, which is better than other  ...  The latest research shows that hyperspectral image classification based on deep convolutional neural network has high accuracy.  ...  The Pavia University data was acquired by the German Reflection Optical Spectral Imager (ROSIS) in Italy in 2003.  ... 
arXiv:2004.12381v2 fatcat:aqlt7j2iy5c7phfkb5nqlij3vm

Hyperspectral image classification using deep convolutional neural networks

Zilong Zhong, Jonathan Li
2017 Journal of Computational Vision and Imaging Systems  
Nevertheless, the high spatial and high spectral dimensionality of each pixel in the hyperspectral imagery hinders the development of hyperspectral image classification.  ...  In this paper, the CNNs have been adopted as an end-to-end pixelwise scheme to classify the pixels of hyperspectral imagery, in which each pixel contains hundreds of continuous spectral bands.  ...  In recent years, deep learning modelswhich include deep belief network (DBN) [3] , auto-encoder (AE) [4] , and convolutional neural network (CNN)have achieved multiple times the highest accuracy in the  ... 
doi:10.15353/vsnl.v3i1.178 fatcat:kptv4lno6be2dlvweomteidqgm

Spectral-Spatial exploration for hyperspectral image classification via the fusion of Fully Convolutional Networks

Liang Zou, Xing liang Zhu, Changfeng Wu, Yong Liu, Lei Qu
2020 IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing  
To address these concerns, we propose a novel spectral-spatial 3-D fully convolutional network (SS3FCN) to jointly explore the spectral-spatial information and the semantic information.  ...  Experimental results on four popular benchmark datasets, including Salinas Valley, Pavia University, Indian Pines, and Houston University, demonstrate that the SS3FCN outperforms state-of-the-art methods  ...  Data Enhancement The combination of more samples and deep network always outperforms the combination of limited sample and shallow network [33] .  ... 
doi:10.1109/jstars.2020.2968179 fatcat:k7yk57pxazdarpkst6wo7ipxzi

Speech Based Depression Detection using Convolution Neural Networks

With the emergence of neural networks and pattern recognition, many researchers have put effort in detecting depression by analysing non-verbal cues, such as facial expressions, gesture, body language  ...  Emotions are a way of expression of one's state of mind in the form of thoughts, feelings or behavioural responses. For a depressed individual, the emotions are often negative in nature.  ...  The emergence of deep neural networks has gained attention in the field of speech analysis as it can learn the features of the input by itself. III.  ... 
doi:10.35940/ijitee.i7076.079920 fatcat:mjnswo4gxrfb3brwepg25j3ztq

IEEE Access Special Section Editorial: Multi-Energy Computed Tomography and its Applications

Hengyong Yu, Yuemin Zhu, Aamir Younis Raja
2021 IEEE Access  
, in 2004, the M.S. degree in physics and applied physics from the University of Engineering & Technology, Lahore, in 2006, and the Ph.D. degree in medical physics from the University of Canterbury, New  ...  In the article ''Image denoising and ring artifacts removal for spectral CT via deep neural network,'' by Lv et al., an image denoising and ring artifacts removal method was proposed via improved fully  ... 
doi:10.1109/access.2021.3105860 fatcat:5tgjmxd6dvgshmfpllxvqofpra

Adipocyte Size Evaluation Based on Photoacoustic Spectral Analysis Combined with Deep Learning Method

Xiang Ma, Meng Cao, Qinghong Shen, Jie Yuan, Ting Feng, Qian Cheng, Xueding Wang, Alexandra Washabaugh, Nicki Baker, Carey Lumeng, Robert O'Rourke
2018 Applied Sciences  
By studying different spectral bands in the entire spectral range using the deep network, a spectral band mostly sensitive to the adipocyte size was identified.  ...  This study proposes an approach of processing the photoacoustic (PA) signal power spectrum using a deep learning method to evaluate adipocyte size in human adipose tissue.  ...  The depth (number of layers) of the deep network largely influences the performance of the network, as illustrated in Figure 4 .  ... 
doi:10.3390/app8112178 fatcat:35wrnmrfnjc6tdoflbyrunfdze

A Novel Method of Hyperspectral Data Classification Based on Transfer Learning and Deep Belief Network

Ke Li, Mingju Wang, Yixin Liu, Nan Yu, Wei Lan
2019 Applied Sciences  
Using the same amount of data, our method based on transfer learning and deep belief network obtains better classification accuracy in a shorter amount of time.  ...  These algorithms require a large amount of data to train the network, while also needing a certain amount of labeled data to fine-tune the network.  ...  Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/app9071379 fatcat:vldpks4iu5atvds4d6lclkf4gy


S. Pu, Y. Song, Y. Chen, Y. Li, J. Zhang, Q. Lin, X. Zhu, Y. Chen, H. Zeng, K. Liao, H. Yu, J. Yuan (+2 others)
2022 ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences  
In particular, graph attention networks (GATs) are good at efficiently processing the graph-structured hyperspectral data by leveraging masked self-attention layers to address the known shortcomings of  ...  and enhancing the expression of local spatial-spectral information.  ...  ACKNOWLEDGEMENTS This work was funded by the Research Funding of the East China University of Technology (No. DHBK2019192).  ... 
doi:10.5194/isprs-annals-v-3-2022-155-2022 fatcat:fg3ta4od4zbhrlvsttq5e4ihym

Advances in Deep Learning for Hyperspectral Image Analysis—Addressing Challenges Arising in Practical Imaging Scenarios [chapter]

Xiong Zhou, Saurabh Prasad
2020 Advances in Computer Vision and Pattern Recognition  
In recent years, there has been an emergence of deep learning algorithms being applied to hyperspectral and multispectral imagery for remote sensing and biomedicine tasks.  ...  Challenges include limited ground truth (annotation is expensive and extensive labeling is often not feasible), and high dimensional nature of the data (each pixel is represented by hundreds of spectral  ...  Deep learning-based methods address this problem in a data-adaptive manner, where the feature learning is undertaken in the context of the overall analysis task in the same network.  ... 
doi:10.1007/978-3-030-38617-7_5 fatcat:23ibk4ojbvepbpikxgjxan4i6e
« Previous Showing results 1 — 15 out of 31,589 results