检索结果-内蒙古大学图书馆

3rd International conference on Digital Society and Intelligent systems, DSInS 2023

作者： Liu, Jiayue Wang, Lei Li, Mingfei Deng, Cong Luo, Weijian Li, Xufeng Lu, Shengzi Zhang, Ruida Chen, Junyang Dong, Shaohua Guangdong Institute of Special Equipment Inspection and Research Guangdong Province Guangzhou510000 China OIL&GAS Pipeline Research Center Beijing102200 China

ISBN: (纸本)9798350331387

para>Defects in pipeline welds are fatal for pipelines, considering that weld negatives need to be electronically preserved due to high preservation costs, easy damage, etc., and that most of the weld defects are judged with the help of intelligent computer recognition algorithms, the assessment of the quality of weld negatives after digitization is critical. At present, domestic and foreign research on the method of weld negative image quality assessment is still immature, especially after the processing of the image details is not clear, and the low quality of the image will directly affect the accuracy of the weld negative defect identification and determination. Therefore, this paper proposes a reference-free image quality assessment algorithm based on analytic sparse representation for this problem, which objectively evaluates the image quality of weld negatives, helps identify and judge weld defects, and has a vital practical role in guaranteeing the regular operation of pipelines. © 2023 IEEE.

关键词： Welds

来源：评论

学校读者我要写书评

暂无评论

Now you see me, now you don't: an exploration of religious exnomination in DALL-E

引用

ETHICS AND INFORMATION TECHNOLOGY 2024年第2期26卷 27-27页

作者： Alfano, Mark Abedin, Ehsan Reimann, Ritsaart Ferreira, Marinus Cheong, Marc Macquarie Univ Dept Philosophy Sydney NSW Australia Univ Melbourne Sch Comp & Informat Syst Melbourne Vic Australia Flinders Univ S Australia Coll Business Govt & Law Adelaide SA Australia

Artificial intelligence (AI) systems are increasingly being used not only to classify and analyze but also to generate images and text. As recent work on the content produced by text and image Generative AIs has shown (e.g., Cheong et al., 2024, Acerbi & Stubbersfield, 2023), there is a risk that harms of representation and bias, already documented in prior AI and natural language processing (NLP) algorithms may also be present in generative models. These harms relate to protected categories such as gender, race, age, and religion. There are several kinds of harms of representation to consider in this context, including stereotyping, lack of recognition, denigration, under-representation, and many others (Crawford in Soundings 41:45-55, 2009;in: Barocas et al., SIGCIS conference, 2017). Whereas the bulk of researchers' attention thus far has been given to stereotyping and denigration, in this study we examine 'exnomination', as conceived by Roland Barthes (1972), of religious groups. Our case study is DALL-E, a tool that generates images from natural language prompts. Using DALL-E mini, we generate images from generic prompts such as "religious person." We then examine whether the generated images are recognizably members of a nominated group. Thus, we assess whether the generated images normalize some religions while neglecting others. We hypothesize that Christianity will be recognizably represented more frequently than other religious groups. Our results partially support this hypothesis but introduce further complexities, which we then explore.

关键词： Artificial intelligence Natural language processing Natural language generation DALL-E Exnomination

来源：评论

学校读者我要写书评

暂无评论

2024 2nd International conference on Computer, Communication and Control, IC4 2024

2024 2nd International Conference on Computer, Communication...

引用

2nd International conference on Computer, Communication and Control, IC4 2024

ISBN: (纸本)9798350387933

The proceedings contain 134 papers. The topics discussed include: an adaptive storage switching algorithm for fault-tolerant network attached storage systems;Covid-19 prediction using machine learning algorithms;energy management of hybrid electric vehicles using cascaded fuzzy logic controller;dynamic lane management with IoT for real-time lane configuration and traffic flow;a closer look at sclera: emerging trends in biometric security;cognitive vision companion: an ai-enhanced support system for the visually impaired;advances in medical image processing for liver tumor recognition: a comprehensive survey;a gradient boosting algorithm to predict energy consumption for home applications;and review on text classification using improved deep learning models.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Automated Bone Segmentation Using a Hybrid Model of DRLSE and Faster R-CNN 25

Automated Bone Segmentation Using a Hybrid Model of DRLSE an...

引用

25th IEEE conference on Signal processing - algorithms, Architectures, Arrangements, and Applications (IEEE SPA)

作者： Naderian, Amirkhashayar Sanei, Saeid Nottingham Trent Univ Dept Comp Sci Nottingham England

ISBN: (数字)9788362065424

ISBN: (纸本)9788362065424

The computer aided diagnosis systems from radiological images has been of interest to researchers mostly for detection of bone fracture or dislocation. The accuracy highly depends on bone segmentation. Any improvement of such systems, particularly for noisy X-ray images, is very valuable. Classical image segmentation depending on image homogeneity are time consuming and require pixel-wise labelling. On the other hand, saliency map based approaches fail to detect the region around the fracture or segment the dislocated bones. In our research we have used transfer learning to train the faster regional convolutional neural network (FCNN) alongside distance regularized level set evolution (DRLSE) to have accurate bone segmentation without any pixel-wise labelling enabling segmentation of the region around the fracture and dislocated bones. We applied the proposed method to a number of hand X-ray images and achieved accuracy values of 95% and average precision-recall of 0.96.

关键词： Bone segmentation distance regularized level set evolution Faster R-CNN image processing medical imaging transfer learning X-ray

来源：评论

学校读者我要写书评

暂无评论

image processing for Knowledge and Comparison of Piranesi’s Carceri Editions 1

引用

3rd International and Interdisciplinary conference on image and Imagination, IMG 2021

作者： Menconero, Sofia Department of History Representation and Restoration of Architecture Sapienza University of Rome Piazza Borghese 9 Rome00186 Italy

ISBN: (数字)9783031259067

ISBN: (纸本)9783031259050

Piranesi printed the 16 etchings of the Carceri d’invenzione in 1761. This version, which is more widespread and better known, derives from the reworking of matrices that the Venetian engraver produced in 1749–50. These are the 14 plates of the Invenzioni capricciose di Carceri all’acquaforte. Many studies have focused on the semantic and formal comparison of the two main versions of the series. Up to now, comparisons have been conducted using juxtaposed pictures, whether original prints or digital reproductions. This procedure involves some limitations due to the effort of synthesis and mental superimposition of the two pictures of different editions to correctly identify the areas where changes occurred. The contribution proposes a method of formal comparison between the two main editions of the series through the digital superimposition of two states of the same plate so that it is easier, more immediate, and more accurate to note the changes. This method is based on the application of image processing algorithms, which let to obtain a synthesis picture in false colours that clearly shows the modified parts of each plate. Observation of the results has suggested the distinction of two main types of modifications carried out by Piranesi (additions/integrations and removal), clearly identifiable in the synthesis pictures. In addition to its implications for Piranesi studies, this research shows how digital image processing can be an essential tool for knowledge in the field of art. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Etching

来源：评论

学校读者我要写书评

暂无评论

Improving The Hiding Capacity of image Steganography with Stego-Analysis

Improving The Hiding Capacity of Image Steganography with St...

引用

2023 IEEE International conference on Integrated Circuits and Communication systems, ICICACS 2023

作者： Grandhe, Padmaja Mallikarjuna Reddy, A. Chillapalli, Kavyasri Koppera, Kavya Thambabathula, Manasa Reddy Surasani, L.P. Pscmrcet Cse Department Vijayawada India Anurag University Department of Cse Hyderabad India

ISBN: (纸本)9798350398465

Communicating online without fearing third-party interventions is becoming a challenge in the modern world. Especially the sectors like the military, and government organizations or private companies sharing sensitive information. They invest a lot of effort and cost into obtaining the advancement of safe communication techniques. image processing encryption techniques using various algorithms promote security over communication channels and using different analysis methods make the tool stand out in providing security to the information. In today's world, there are various steganographic mechanisms that convert the secret message into stego medium and send it across various communication channels. Using algorithms like Blind Hide promotes the security of the message along with using multiple analysis methods that will further improve the tool in giving out information of encoded accuracy, size of stego of the secret message. The aim is to generate a tool that will give out a benchmark value of how precisely the message is stored in the cover file. Using Stego and bulk analysis the information about the presence of the stego medium in the message can be known to the user. All these analysis methods make the tool more enhanced and secure. © 2023 IEEE.

关键词： Steganography

来源：评论

学校读者我要写书评

暂无评论

BCD-Unet:A Novel Water Areas Segmentation Structure for Remote Sensing image 7

BCD-Unet:A Novel Water Areas Segmentation Structure for Remo...

引用

7th International conference on Intelligent Computing and Control systems, ICICCS 2023

作者： Jonnala, Naga Surekha Gupta, Neha Vasantrao, Chafle Pratiksha Mishra, Anoop Kumar VIT-AP University School of Electronics Engineering Amaravati India

ISBN: (纸本)9798350397253

Surface water resource identification is one of the main techniques used in remote sensing image analysis. This is necessary to stop calamities like floods and droughts. Feature selection based on prior information and complex spectrum analysis are required by machine learning algorithms for extracting waterbodies. The use of deep learning-based algorithms, which have just lately been considered, has eliminated the need for manually extracting features, but they require a large amount of training data and processing power to achieve outstanding performance. Because of this, each deep learning architecture that has been demonstrated may recognize some of the patterns that are present in the given situations. To develop a reliable model for identifying surface water, this research trains and optimises deep learning architecture that have been presented in different disciplines using surface water data. This is accomplished by the model's output more accurately by extracting the finalized segmented mask of water bodies. Our suggested approach was tested using a Kaggle dataset. This dataset included Sentine1-2 satellite images of numerous water bodies. The suggested method outperforms current best practises in semantic segmentation performance. © 2023 IEEE.

关键词： Surface waters

来源：评论

学校读者我要写书评

暂无评论

Robust Non-convex Model-Based Approach for Deep Learning-Based image processing 4th

Robust Non-convex Model-Based Approach for Deep Learning-Ba...

引用

4th International conference on Numerical Computations: Theory and algorithms, NUMTA 2023

作者： Morotti, Elena Evangelista, Davide Loli Piccolomini, Elena Department of Political and Social Sciences University of Bologna Bologna Italy Department of Mathematics University of Bologna Bologna Italy Department of Computer Science and Engineering University of Bologna Bologna Italy

ISBN: (纸本)9783031812408

Total p-norm Variation (TpV) is a well-established technique in image processing, used to denoise and preserve edges. However, the related non-convex minimization is still a challenging task in optimization, both for the computational cost and for convergence guarantees toward a good local minimum. We propose a framework, called (TpV)2, embedding a convolutional neural network to speed up the iterative reconstruction while preserving converging features. The resulting hybrid method is robust and accurate. Verifications and comparisons illustrate that the proposed method is effective and promising. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Inverse Problems Neural Network Non-convex optimization

来源：评论

学校读者我要写书评

暂无评论

Increasing the accuracy of object boundaries detection and the quality of machine vision system data in terms of machine processing by a collaborative complex 32

Increasing the accuracy of object boundaries detection and t...

引用

conference on Signal processing, Sensor/Information Fusion, and Target Recognition XXxii

作者： Semenishchev, Evgenii Shraifel, Igor Makov, Sergey Surkov, Egor Zelensky, Aleksandr Voronin, Viacheslav Tula State Univ TulSU 92 Sq Lenina Tula 300012 Tula Russia Don State Tech Univ DSTU Inst Serv & Entrepreneurship 147 St Shevchenko Shakhty 346500 Rostov Russia Moscow State Tech Univ STANKIN 1a Vadkovsky Moscow 127055 Russia

ISBN: (数字)9781510662117

ISBN: (纸本)9781510662100;9781510662117

To article proposes an approach to improve the quality of data used in various processes based on machine vision systems. The paper proposes a combined approach to applying the method of multi-criteria processing based on the use of a combined criterion in order to implement an edge detector, smoothing and separation areas of the background / object in the image. The application of the method allows eliminating the noise caused by external factors (such as dust and water suspension on the lens or space). The generated data make it possible to form an adaptive criterion for changing the correction parameters for a non-linear change in color balance in areas of increased detail or selected masks of changes blocks. The proposed algorithms make it possible to increase the visibility of small elements, reduce the noise component, while maintaining the boundaries of objects, increase the accuracy of selecting the boundaries of objects and the visual quality of data. As test data used to evaluate the effectiveness, nature data and expert evaluation results for test images obtained by a machine vision system with a sensor with a resolution of 1024x768 (8-bit, color image, visible range) are used. images of simple shapes are used as analyzed objects.

关键词： boundaries detection image machine vision image smoothing

来源：评论

学校读者我要写书评

暂无评论

CROMA: Remote Sensing Representations with Contrastive Radar-Optical Masked Autoencoders 37

CROMA: Remote Sensing Representations with Contrastive Radar...

引用

37th conference on Neural Information processing systems (NeurIPS)

作者： Fuller, Anthony Millard, Koreen Green, James R. Carleton Univ Dept Syst & Comp Engn Ottawa ON Canada Carleton Univ Dept Geog & Environm Studies Ottawa ON Canada

ISBN: (纸本)9781713899921

A vital and rapidly growing application, remote sensing offers vast yet sparsely labeled, spatially aligned multimodal data;this makes self-supervised learning algorithms invaluable. We present CROMA: a framework that combines contrastive and reconstruction self-supervised objectives to learn rich unimodal and multimodal representations. Our method separately encodes masked-out multispectral optical and synthetic aperture radar samples-aligned in space and time-and performs cross-modal contrastive learning. Another encoder fuses these sensors, producing joint multimodal encodings that are used to predict the masked patches via a lightweight decoder. We show that these objectives are complementary when leveraged on spatially aligned multimodal data. We also introduce X- and 2D-ALiBi, which spatially biases our cross- and self-attention matrices. These strategies improve representations and allow our models to effectively extrapolate to images up to 17.6x larger at test-time. CROMA outperforms the current SoTA multispectral model, evaluated on: four classification benchmarks-finetuning (*** arrow 1.8%), linear (*** arrow 2.4%) and nonlinear (*** arrow 1.4%) probing, kNN classification (*** arrow 3.5%), and K-means clustering (*** arrow 8.4%);and three segmentation benchmarks (*** arrow 6.4%). CROMA's rich, optionally multimodal representations can be widely leveraged across remote sensing applications.

关键词： image enhancement

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：