检索结果-内蒙古大学图书馆

Improving u-net architecture and graph cuts optimization to classify arterioles and venules in retina fundus images

JOuRNAL OF INTELLIGENT & FuZZY SYSTEMS 2022年第4期42卷 4015-4026页

作者： Nguyen Thanh Binh Nguyen Mong Hien Dang Thanh Tin Ho Chi Minh City Univ Technol HCMUT Fac Comp Sci & Engn Dept Informat Syst Ho Chi Minh City Vietnam Vietnam Natl Univ Ho Chi Minh City Linh Trung Ward Ho Chi Minh City Vietnam Tra Vinh Univ Tra Vinh Vietnam Ho Chi Minh City Univ Technol HCMUT Fac Elect & Elect Engn Informat Syst Engn Lab Ho Chi Minh City Vietnam

The central retinal artery and its branches supply blood to the inner retina. Vascular manifestations in the retina indirectly reflect the vascular changes and damage in organs such as the heart, kidneys, and brain because of the similar vascular structure of these organs. The diabetic retinopathy and risk of stroke are caused by increased venular caliber. The degrees of these diseases depend on the changes of arterioles and venules. The ratio between the calibers of arterioles and venules (AVR) is various. AVR is considered as the useful diagnostic indicator of different associated health problems. However, the task is not easy because of the lack of information of the features being used to classify the retinal vessels as arterioles and venules. This paper proposed a method to classify the retinal vessels into the arterioles and venules based on improving u-net architecture and graph cuts. The accuracy of the proposed method is about 97.6%. The results of the proposed method are better than the other methods in RITE dataset and AVRDB dataset.

关键词： Arterioles venules u-net architecture graph cuts retinal blood vessels

来源：评论

学校读者我要写书评

暂无评论

2D Fluid Flows Prediction Based on u-net architecture 5

2D Fluid Flows Prediction Based on U-Net Architecture

引用

5th International Conference on Artificial Intelligence in Information and Communication (ICAIIC)

作者： Prihatno, Aji Teguh Kang, Hyoeun Choi, Chang Woo Le, Thi-Thu-Huong Heo, Shinwook Kim, Myeongkil Kim, Howon Pusan Natl Univ Comp Sci & Engn Busan 609735 South Korea Pusan Natl Univ Blockchain Platform Res Ctr Busan 609735 South Korea SmartM2M Busan 48058 South Korea

ISBN: (纸本)9781665456456

Computational fluid dynamics (CFD) solvers provide helpful components and amenities for industrial development and the advancement of fluid flow simulations. Nevertheless, CFD solvers are not advantageous since iterative simulations require high computational resources and enormous memory for complex calculations. The deep neural network-based CFD data-driven learning method eliminates these constraints by lowering the compensation between model complicatedness and precision. In this paper, we present the feasibility of predicting fluid flow velocity fields based on CFD using u-net architecture, a subdomain of deep learning. The experimental results show that the u-net architecture can predict fluid flow with a total loss of 0.2223, a validation loss of 0.2728, and an accuracy of 86% from our private dataset. Our u-net model can be used to predict fluid flows, which has been proven.

关键词： CFD Laminar Flow Deep Learning Convolutional Neural networks u-net architecture

来源：评论

学校读者我要写书评

暂无评论

Semantic Segmentation using u-net architecture for Change Detection on Hyperspectral Imagery

Semantic Segmentation Using U-Net Architecture for Change De...

引用

2023 International Conference on Sustainable Communication networks and Application, ICSCNA 2023

作者： Bidari, Indira Chickerur, Satyadhyan Kadam, Srishti School of Computer Science & Engineering Kle Technological University Karnataka Hubballi India

ISBN: (纸本)9798350313987

Modern research focuses on the creation of cutting-edge image processing methods designed for the analysis of vast amounts of spatial and temporal data gathered from satellite imagery. These techniques have a primary objective to enhance the clarity and precision of satellite images, consequently unveiling insights into previously undiscovered changes in land cover. Recent strides in image processing, with a particular focus on the utilization of Fully Convolutional Neural networks (FCNNs) have become essential tools. The primary goal is to evaluate the performance of the u-net architecture in recognizing changes in multi-temporal satellite images and associating objects inside land-covered regions. Multi-Class Change Detection (MCD) is expanded upon by Semantic Change Detection (SCD). It holds a central role in this endeavor, offering detailed insights into modifications occurring in specific geographic locations, as well as changes in land cover and land use. This high-fidelity change data holds wide-ranging applications and significantly enriches our comprehension of dynamic environmental shifts. © 2023 IEEE.

关键词： Hyperspectral Image Satellite Imagery Semantic Change Detection u-net architecture

来源：评论

学校读者我要写书评

暂无评论

RIBM3Du-net: Glioma tumour substructures segmentation in magnetic resonance images using residual-inception block with modified 3D u-net architecture

引用

INTERNATIONAL JOuRNAL OF IMAGING SYSTEMS AND TECHNOLOGY 2024年第2期34卷 e23056-e23056页

作者： Shajahan, Syedsafi Pathmanaban, Sriramakrishnan Tiruvenkadam, Kalaiselvi Kalasalingam Acad Res & Educ Deemed Univ Sch Comp Dept Comp Applicat Krishnankoil India Amrita Vishwa Vidyapeetham Deemed Univ Amrita Sch Phys Sci Dept Math Coimbatore India Gandhigram Rural Inst Deemed Univ Dept Comp Sci & Applicat Dindigul India Amrita Vishwa Vidyapeetham Deemed Univ Amrita Sch Phys Sci Dept Math Coimbatore 641112 Tamil Nadu India

Glioma brain tumour is one of the life-threatening diseases in the world. Tumour substructure segmentation by physicians is a time-consuming task with the magnetic resonance imaging (MRI) technique due to the size of clinical data. An automatic and well-trained method is essential to detect and segment the tumour which increase the survival of the patients. The proposed work aims to produce high accuracy on glioma substructures segmentation with less computation time using deep learning. From the literature survey, the following challenges are found: (i) computing complex spatial boundaries between normal and tumour tissues, (ii) feature reduction and (iii) overfitting problems. Hence, we proposed a fully automatic glioma tumour segmentation using a residual-inception block (RIB) with a modified 3D u-net (RIBM3Du-net). It includes three phases: pre-processing, modified 3D u-net segmentation and post-processing. From the results, RIB with u-net enhances the segmentation accuracy. GPu parallel architecture reduces the computation time while training and testing. For quantitative analysis, comprehensive experiment results were computed and compared with state-of-the-art methods. It achieves better Dice scores on enhancing tumour, tumour core and complete tumour of 87%, 87% and 94%, respectively. GPu speedup folds yield up to 48x when compared with CPu. Quantitatively, 3D glioma volume is rendered from the obtained segmented results and estimated using the Cavalieri estimator.

关键词： deep learning glioma segmentation graphics processing units residual-inception block u-net architecture

来源：评论

学校读者我要写书评

暂无评论

A u-net architecture as a surrogate model combined with a geostatistical spectral algorithm for transient groundwater flow inverse problems

引用

ADVANCES IN WATER RESOuRCES 2024年 189卷

作者： Lauzon, Dany Polytech Montreal Dept Genies Civil Geol & Mines POB 6079 Succ Ctr Ville Montreal PQ H3C 3A7 Canada

Characterizing groundwater flow parameters is crucial for understanding complex aquifer systems, and inverse techniques play a fundamental role in modeling hydrogeological parameters and assessing their uncertainties. Nonetheless, the use of a forward model in these methods can be highly time-consuming, especially with an increasing number of model parameters. To address this issue, we propose a surrogate model based on a u -net architecture that replaces the transient groundwater flow model, reducing runtime and enabling a fast quantification of uncertainties related to key parameters, including heterogeneous hydraulic conductivity, boundary conditions, specific storage, and pumping rate. The surrogate is trained using limited evaluations of the forward model to learn the physical relationship between hydraulic conductivity fields and transient hydraulic heads measured on -site. The physical principles of the studied problem, including boundary conditions, specific storage, and source terms, are also mapped and introduced as inputs to the model to enhance its understanding of the governing equation of transient groundwater flow. To speed up learning using image-image regression, the previously predicted transient hydraulic heads also serve as an input to predict the transient heads at the current time step. Once the model is trained, we use a spectral geostatistical method to solve the inverse problem, a pumping test of 12 h, using the surrogate model in place of the forward model. Our study demonstrates that the trained u -net accurately reproduces the state variables corresponding to a specific parameter field, and in terms of computational demand, using u -net as a surrogate model reduces the required computational time by approximately an order of magnitude for the defined problem. The proposed approach offers an efficient and accurate method for groundwater flow parameter characterization and uncertainty quantification in complex aquifer systems.

关键词： Surrogate modeling Inverse problem Stochastic hydrogeology Applied geostatistics u-net architecture Pumping test

来源：评论

学校读者我要写书评

暂无评论

Segmentation of Sella Turcica in X-ray Image based on u-net architecture

引用

Procedia Computer Science 2023年 218卷 828-835页

作者： Kaushlesh Singh Shakya K Priti Manojkumar Jaiswal Amit Laddi Academy of Scientific & Innovative Research (AcSIR) Ghaziabad 201002 India CSIR-Central Scientific Instruments Organisation Chandigarh 160030 India Oral Health Sciences (Dental) Centre PGIMER Chandigarh 160012 India

The computer-aided evaluation of medical imaging and orthodontic examinations requires the segmentation of Sella Turcica for feature extraction using cephalometric X-ray images, which is a novel and complex task. However, x-ray image segmentation is more complicated than CT and MR images due to less association of dense thin white tissue of Sella Turcica that is difficult to distinguish from cephalometric radiographs. This paper presents deep learning methods for completely automated Sella Turcica segmentation from cephalometric X-ray images using a fully connected neural network ("u-net"). The experiments were performed and analyzed using secondary datasets provided by PGIMER, Chandigarh. The different pre-trained u-net models (VGG16 and 19, Resnet34, 50, and 152) are used as an encoder to improve the prediction accuracy of the Sella structure. The author compared the Iou score of all the considered pre-trained models and found that VGG19 and Resnet34 have high Iou scores compared to other models, but out of all the considered models, Resnet34 gives good prediction results.

关键词： Sella Turcica X-ray image Segmentation u-net architecture Deep learning

来源：评论

学校读者我要写书评

暂无评论

A Low-light Image Enhancement Method Based on Improved u-net architecture for Nighttime Traffic Scene Image Recovery

A Low-light Image Enhancement Method Based on Improved U-Net...

引用

2024 世界交通运输大会

作者： Xiaohang Jia Shujing Fang Sai Zhou Yangyang Li School of Information Engineering Chang'an University

Images obtained in extreme dark light scenes usually suffer from severe quality degradation,such as reduced contrast and loss of content and color,leading to a perceptual decline in visual quality,which adversely affects intelligent traffic surveillance based on video *** deep learning-based low-light enhancement method provides an effective solution for the quality restoration of nighttime traffic ***,the low-frequency luminance and high-frequency texture information of low-light images are coupled with each other,which is difficult to take into account by most enhancement methods,resulting in luminance imbalance and ***,accurate color restoration is rarely considered during network optimization,leading to color distortion in *** address this,we propose a low-light image enhancement method based on an improved u-net *** design a high-low frequency decoupling multi-branch convolutional block(MBC block)to simultaneously focus on extracting high and low-frequency features of images and utilize spatial and channel attention to further enhance network representation *** model optimization,we introduce a color fidelity loss,optimizing the brightness and color of images by decomposing RGB color values into vector lengths and angles,to achieve more natural color *** on a low-lightenhanced dataset show that our method balances image brightness and high-frequency detail recovery with more accurate color *** also shows excellent enhancement results on the nighttime traffic image recovery task,which contributes to the stable operation of a real-time traffic monitoring system.

关键词： Nighttime traffic image Low-light enhancement u-net architecture Multi-branch convolution Spatial and channel attention Color fidelity loss

来源：评论

学校读者我要写书评

暂无评论

Improvement of u-net architecture for Image Binarization with Activation Functions Replacement 13

Improvement of U-Net Architecture for Image Binarization wit...

引用

13th International Conference on Machine Vision

作者： Gayer, Alexander, V Sheshkus, Alexander, V Nikolaev, Dmitri P. Arlazarov, Vladimir V. Smart Engines Serv LLC Moscow Russia Natl Univ Sci & Technol MISIS Moscow Russia Russian Acad Sci Fed Res Ctr Comp Sci & Control Moscow Russia Moscow Inst Phys & Technol Moscow Russia RAS Kharkevich Inst Inst Informat Transmiss Problems Moscow Russia

ISBN: (纸本)9781510640412

In this work we study the effect of activation functions in a neural network. We consider how activation functions with different properties and their combination affect the final quality of the model. Due to optimization and speed performance issues with most of bounded functions that are represented by sigmoids, we propose the generalized version of SoftSign function - ratio function (rf). Its shape greatly depends on introduced degree parameter, which in theory leads to new interesting property - contraction to zero. For evaluation, we chose image binarization problem: based on u-net architecture of DIBCO-2017 winners, we conducted all experiments with replacing activation functions only. Our research has led us to the state-of-the-art results in binarization quality on DIBCO-2017 test dataset. u-net with modified activation functions significantly outperforms all existing solutions in all metrics.

关键词： activation function deep learning u-net architecture binarization DIBCO

来源：评论

学校读者我要写书评

暂无评论

Towards speech enhancement using a variational u-net architecture 29

Towards speech enhancement using a variational U-Net archite...

引用

29th European Signal Processing Conference (EuSIPCO)

作者： Nustede, Eike J. Anemueller, Joern Carl von Ossietzky Univ Oldenburg Computat Audit Grp Dept Med Phys & Acoust Oldenburg Germany Carl von Ossietzky Univ Oldenburg Computat Audit Grp Cluster Excellence Hearing4all Oldenburg Germany

ISBN: (纸本)9789082797060

We investigate the viability of a variational u-net architecture for denoising of single-channel audio data. Deep network speech enhancement systems commonly aim to estimate filter masks, or opt to work on the waveform signal, potentially neglecting relationships across higher dimensional spectrotemporal features. We study the adoption of a probabilistic (variational) bottleneck into the classic u-net architecture for direct spectral reconstruction. Evaluation of several ablation network variants is carried out using signal-to-distortion ratio and perceptual measures, on audio data that includes known and unknown noise types as well as reverberation. Our experiments show that the residual (skip) connections in the proposed system are a prerequisite for successful spectral reconstruction, i.e., without filter mask estimation. Results show, on average, an advantage of the proposed variational u-net architecture over its classic, non-variational version in signal enhancement performance under reverberant conditions of 0.31 and 6.98 in PESQ and STOI scores, respectively. Anecdotal evidence points to improved suppression of impulsive noise sources with the variational u-net compared to the recurrent mask estimation network baseline.

关键词： Speech enhancement u-net architecture variational autoencoder deep learning audio source separation

来源：评论

学校读者我要写书评

暂无评论

Sperm Cell Segmentation in Digital Micrographs based on Convolutional Neural networks using u-net architecture 34

Sperm Cell Segmentation in Digital Micrographs based on Conv...

引用

34th IEEE International Symposium on Computer-Based Medical Systems (IEEE CBMS)

作者： Melendez, Roy Beltran Castanon, Cesar Medina-Rodriguez, Rosario Pontificia Univ Catolica Peru Postgrad Sch Dept Engn Av Univ 1801 Lima Peru

ISBN: (纸本)9781665441216

Human infertility is considered a serious disease of the reproductive system that affects more than 10% of couples worldwide, and more than 30% of reported cases are related to men. The crucial step in evaluating male infertility is a semen analysis, highly dependent on sperm morphology. However, this analysis is done at the laboratory manually and depends mainly on the doctor's experience. Besides, it is laborious, and there is also a high degree of interlaboratory variability in the results. This article proposes applying a specialized convolutional neural network architecture (u-net), which focuses on the segmentation of sperm cells in micrographs to overcome these problems. The results showed high scores for the model segmentation metrics such as precision (93%), Iou score (88%), and DICE score of 94%. Moreover, we can conclude that u-net architecture turned out to be a good option to carry out the segmentation of sperm cells.

关键词： sperm cell micrographs image segmentation u-net architecture deep learning

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：