检索结果-内蒙古大学图书馆

Efficient primal-dual fixed point algorithms with dynamic stepsize for composite convex optimization problems

MULTIDIMENSIONAL systems AND SIGNAL processing 2019年第3期30卷 1531-1544页

作者： Wen, Meng Tang, Yuchao Cui, Angang Peng, Jigen Xian Polytech Univ Sch Sci Xian 710048 Shaanxi Peoples R China Guangzhou Univ Sch Math & Informat Sci Guangzhou 510006 Guangdong Peoples R China NanChang Univ Dept Math Nanchang 330031 Jiangxi Peoples R China Xi An Jiao Tong Univ Sch Math & Stat Xian 710049 Shaanxi Peoples R China

In this paper, we propose a new primal-dual fixed point algorithm with dynamic stepsize (PDFP2ODSn) for solving convex minimization problems involving the sum of a smooth function with Lipschitzian gradient and the composition of a nonsmooth convex function with a continuous linear operator. Based on modified Mann iteration and the firmly nonexpansive properties of the proximity operator, we achieve the convergence of the proposed algorithm. Moreover, we give the connection of the proposed algorithm with other existing PDFP2O (Chen et al. in Inverse Probl 29:025011-025033, 2013). Finally, we illustrate the efficiency of PDFP2ODSn through some numerical examples on the CT image reconstruction problem. Numerical results show that our iterative algorithm (PDFP2ODS) performs better than the original one (PDFP2O).

关键词： Fixed point algorithm Convex separable minimization Proximity operator Dynamic stepsize

来源：评论

学校读者我要写书评

暂无评论

Skin cancer classification computer system development with deep learning 10

Skin cancer classification computer system development with ...

引用

Workshop on Information Technology and Scientific Computing in the Framework of the 10th International conference Information and Telecommunication Technologies and Mathematical Modeling of High-Tech systems, ITTMM 2020

作者： Demidova, Anastasia V. Kulyabov, Dmitry S. Shchetinin, Eugeny Yu. Department of Applied Probability and Informatics Peoples' Friendship University of Russia 6 Miklukho-Maklaya St. Moscow117198 Russia Laboratory of Information Technologies Joint Institute for Nuclear Research 6 Joliot-Curie St. Dubna Moscow region141980 Russia Department of Data Analysis Decision Making and Financial Technologies Financial University under the Government of the Russian Federation 49 Leningradsky pr. Moscow111234 Russia

Melanoma is a deadly form of skin cancer that is often undiagnosed or misdiagnosed as a benign skin lesion. Its early detection is extremely important, since the life of patients with melanoma depends on accurate and early diagnosis of the disease. However, doctors often rely on personal experience and assess each patient's injuries based on a personal examination. Clinical studies allow us to get the accuracy of the diagnosis of melatoma from 65 to 80 percents, which was a good result for some time. However, modern research claims that the use of dermoscopic images in diagnosis significantly increases the accuracy of diagnosis of skin lesions. The visual differences between melanoma and benign skin lesions can be very small, making diagnosis difficult even for an expert doctor. Recent advances in the use of artificial intelligence methods in the analysis of medical images have made it possible to consider the development of intelligent medical diagnostic systems based on visualization as a very promising direction that will help the doctor in making more effective decisions about the health of patients and making a diagnosis at an early stage and in adverse conditions. In this paper, we propose an approach to solving the problem of classification of skin diseases, namely, melanoma at an early stage, based on deep learning. In particular, a solution to the problem of classification of a dermoscopic image containing either malignant or benign skin lesions is proposed. For this purpose, the deep neural network architecture was developed and applied to image processing. Computer experiments on the ISIC data set have shown that the proposed approach provides 92% accuracy on the test sample, which is significantly higher than other algorithms in this data set have shown. © 2020 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). CEUR Workshop Proceedings (***)

关键词： image classification

来源：评论

学校读者我要写书评

暂无评论

An enhanced non local mean method for hole filling in the depth image-based rendering system 11

An enhanced non local mean method for hole filling in the de...

引用

11th International conference on Digital image processing (ICDIP)

作者： Zhang, Zihao Wang, Yuanqing Zhan, Lingli Nanjing Univ Sch Elect Sci & Engn Nanjing 210046 Peoples R China Nanjing Univ Key Lab Intelligent Opt Sensing & Manipulat Minist Educ Nanjing 210046 Peoples R China

ISBN: (数字)9781510630765

ISBN: (纸本)9781510630765

Depth image based rendering (DIBR) is the most widely used technology among synthesis algorithms. Hole filling is a challenge in producing desirable synthesized images. In this paper, we propose an enhanced non local mean based hole filling method. Color, gradient and depth information is combined to select the optimal candidate patches. The missing information from holes is then formed by aggregating multiple candidate patches. Furthermore, an efficient invalid pixel classification method based on their chararcteristics is proposed to divide invalid pixels into three types, and use different methods to fill them, and reduce the computational load of the hole filling unit. The results show that the proposed method has a better robustness and performance for hole filling in DIBR systems than other hole filling based on algorithms.

关键词： Depth image based rendering(DIBR) Hole filling Non local mean based hole filling method

来源：评论

学校读者我要写书评

暂无评论

Joint unmixing-deconvolution algorithms for hyperspectral images 27

Joint unmixing-deconvolution algorithms for hyperspectral im...

引用

27th European Signal processing conference (EUSIPCO)

作者： Song, Yingying Djermoune, El-Hadi Brie, David Richard, Cedric Univ Lorraine CRAN CNRS Vandoeuvre Les Nancy France Univ Cote dAzur CNRS OCA Nice France

ISBN: (纸本)9789082797039

This paper combines supervised linear unmixing and deconvolution problems to increase the resolution of the abundance maps for industrial imaging systems. The joint unmixing-deconvolution (JUD) algorithm is introduced based on the Tikhonov regularization criterion for offline processing. In order to meet the needs of industrial applications, the proposed JUD algorithm is then extended for online processing by using a block Tikhonov criterion. The performance of JUD is increased by adding a non-negativity constraint which is implemented in a fast way using the quadratic penalty method and fast Fourier transform. The proposed algorithm is then assessed using both simulated and real hyperspectral images.

关键词： Hyperspectral image unmixing hyperspectral image deconvolution non-negative Tikhonov regularization

来源：评论

学校读者我要写书评

暂无评论

Comparison of three denoising methods for cardiac diffusion tensor imaging

Comparison of three denoising methods for cardiac diffusion ...

引用

International conference on Mechanical, Control and Computer Engineering (ICMCCE)

作者： xi Xu Shengping Liu Muzi Guo Yanjie Zhu Rui Chen Yuxin Yang Chongqing University of Technology Chongqing China Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen China Guangdong Provincial People's Hospital Guangdong Academy of Medical Sciences Guangzhou China

ISBN: (纸本)9781665423144;9781665446235

Cardiac diffusion tensor imaging (CDTI) provides unique information on the structure, organization, and integrity of the myocardium. However, the inherently low signal-to-noise ratio (SNR) limits the accuracy in parameter estimations. Recently, several post-processing algorithms have been proposed to improve image quality in brain DTI, but their performance is unknown in CDTI. This paper evaluates three different denoising algorithms (ANLM, LPCA, and MPPCA) in CDTI regarding image quality and accuracy of parameters with simulation and ex-vivo experiments. We add white Gaussian noise to the simulated DWI images and average ex-vivo data with various repetitions to imitate CDTI images with different SNRs. Then, denoising methods are applied, and the performance is evaluated by SNR and RMSE. The results show that the SNR is improved by 72.38%, 127.01%, 0.6% in simulation, and 327.38%, 474.56%, 173.19% in ex-vivo experiments by the three methods, respectively. The uncertainty of parameters is reduced by all three algorithms in the ex-vivo experiment.

关键词： image quality Uncertainty Parameter estimation Gaussian noise Noise reduction Process control Organizations

来源：评论

学校读者我要写书评

暂无评论

INDUCTIVE CONFORMAL PREDICTOR FOR SPARSE CODING CLASSIFIERS: APPLICATIONS TO image CLASSIFICATION 44

INDUCTIVE CONFORMAL PREDICTOR FOR SPARSE CODING CLASSIFIERS:...

引用

44th IEEE International conference on Acoustics, Speech and Signal processing (ICASSP)

作者： Matiz, Sergio Barner, Kenneth E. Univ Delaware Dept Elect & Comp Engn Newark DE 19716 USA

ISBN: (纸本)9781479981311

Conformal prediction uses the degree of strangeness (nonconformity) of new data instances to determine the confidence values of new predictions. We propose an inductive conformal predictor for sparse coding classifiers, referred to as ICP-SCC. Our contribution is twofold: first, we present two nonconformity measures that produce reliable confidence values;second, we propose a batch mode active learning algorithm within the conformal prediction framework to improve classification performance by selecting training instances based on two criteria, informativeness and diversity. Experiments conducted on face and object recognition databases demonstrate that ICP-SCC improves the classification accuracy of state-of-the-art dictionary learning algorithms while producing reliable confidence values.

关键词： Conformal prediction sparse coding dictionary learning image classification active learning

来源：评论

学校读者我要写书评

暂无评论

An Approach to Defense Dictionary Attack with Message Digest Using image Salt 13th

An Approach to Defense Dictionary Attack with Message Digest...

引用

13th International conference on Ubiquitous Information Management and Communication (IMCOM)

作者： Park, Sun-young Kim, Keecheon Konkuk Univ Dept IT Convergence Informat Secur Seoul South Korea Konkuk Univ Dept Comp Sci & Engn Seoul South Korea

ISBN: (纸本)9783030190637;9783030190620

Hash algorithms have been widely used for cryptography. It has been impossible to decrypt the ciphertexts generated through hash algorithms, as an operation that damages the original text is performed. However, various methods of attack occurred over time after the algorithm was developed. The vulnerability of SHA1 (an old hash algorithm) has been revealed, and there has been a great deal of data available for dictionary attacks. Although the industry has been gradually refraining from using SHA1, it remains in use in some existing systems for various reasons. For example, when problems resulting from service interruption or mass update are critical, updating the encryption algorithm can be a burden. In this study, we aim to increase the complexity of ciphertexts by postprocessing hash ciphertext. For that, image salting techniques are used using two-dimensional array masking. This will allow the use of hash ciphertexts with increased complexity in some devices that are forced to use old hash algorithms for various reasons.

关键词： Hash algorithm image processing Rainbow table Message digest Two-dimensional array image salting

来源：评论

学校读者我要写书评

暂无评论

Variational Denoising Network: Toward Blind Noise Modeling and Removal 33

Variational Denoising Network: Toward Blind Noise Modeling a...

引用

33rd conference on Neural Information processing systems (NeurIPS)

作者： Yue, Zongsheng Yong, Hongwei Zhao, Qian Zhang, Lei Meng, Deyu Xi An Jiao Tong Univ Sch Math & Stat Xian Shaanxi Peoples R China Hong Kong Polytech Univ Dept Comp Kowloon Hong Kong Peoples R China Alibaba Grp DAMO Acad Shenzhen Peoples R China Macau Univ Sci & Technol Fac Informat Technol Macau Peoples R China

Blind image denoising is an important yet very challenging problem in computer vision due to the complicated acquisition process of real images. In this work we propose a new variational inference method, which integrates both noise estimation and image denoising into a unique Bayesian framework, for blind image denoising. Specifically, an approximate posterior, parameterized by deep neural networks, is presented by taking the intrinsic clean image and noise variances as latent variables conditioned on the input noisy image. This posterior provides explicit parametric forms for all its involved hyper-parameters, and thus can be easily implemented for blind image denoising with automatic noise estimation for the test noisy image. On one hand, as other data-driven deep learning methods, our method, namely variational denoising network (VDN), can perform denoising efficiently due to its explicit form of posterior expression. On the other hand, VDN inherits the advantages of traditional model-driven approaches, especially the good generalization capability of generative models. VDN has good interpretability and can be flexibly utilized to estimate and remove complicated non-i.i.d. noise collected in real scenarios. Comprehensive experiments are performed to substantiate the superiority of our method in blind image denoising.

关键词： image denoising

来源：评论

学校读者我要写书评

暂无评论

Digital Signal processing Based on Spline Functions

Digital Signal Processing Based on Spline Functions

引用

2019 International conference on Information Science and Communications Technologies, ICISCT 2019

作者： Mirzaev, Avaz Khalilov, Sirojiddin Information Technologies Dept Tashkent Uzbekistan

ISBN: (纸本)9781728125640

The development of the modern world's information and communication technologies and their implementation in various industries are important, are essential to the creation of parallel algorithms for solving problems of the recovery and digital signal processing, on the basis of the treatment processes a multi-core architecture and the search for optimal solutions. The current stages of development of the structure of machines, complexes and systems installed at facilities that are part of the structure of devices specialized for conducting scientific research or located on rolling stock are characterized by on-line analysis of complex processes and fields and increased demands on the speed of processing large amounts of data in real time. Particular attention is paid to the development of methods for solving problems of image and signal processing by using modern methods, algorithms and structural tools, architecture of computing and software. In that work the result of the study of methods of approximation of functions and functional dependencies obtained in the course of experiments for the selected class of signals improved method of spline functions on the basis of analytical analysis, parallel algorithms for digital signal processing have been developed, efficient algorithms have been developed for digital processing of vibrio signals in multi-core processors based on cubic and bi cubic basis splines. © 2019 IEEE.

关键词： Digital signal processing

来源：评论

学校读者我要写书评

暂无评论

Identification of Drowning Victims in Freshwater Bodies using Drift Prediction and image processing based on Deep Learning

Identification of Drowning Victims in Freshwater Bodies usin...

引用

International conference on Advanced Computer Control, ICACC

作者： Anjana Unnikrishnan A T Roshni P R Anusha Anju M Vinny C K Anuraj Toc H Institute of Science & Technology Ernakulam India

ISBN: (纸本)9781665439206

Year after year drowning deaths are increasing tremendously, making it the 3rd leading cause of unintentional injury deaths worldwide. Drift prediction methodology is typically not used in river ecosystems and conventional methods for human rescue do not account for feasible and faster human detection. Utilization of multiple sensor data in underwater human rescue applications can capacitate faster human detection. This paper discusses the design, implementation, and testing of such an underwater human detection system, which spots the victim drifting or drowning in freshwater ecosystems. The water flow sensor attached to this portable device can calculate drift distance to track down the victim. The ultrasonic sensor activates the underwater camera upon detecting an object, to facilitate real-time human localization. We performed real-time object detection on a custom dataset by applying DarkNet-53 pre-trained weights on YOLOv3 architecture and a mean Average Precision (mAP) of 98.0% was achieved. The system attained a detection depth of 5m. Combined action of drift distance calculator and YOLOv3 real-time detection model can speed up underwater human extrication.

关键词： Performance evaluation Location awareness Ecosystems Object detection Phase frequency detectors Prediction algorithms Real-time systems

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：