检索结果-内蒙古大学图书馆

Intelligent Systems conference

作者： Park, Jiyoon Branksome Hall Asia Seogwipo South Korea

ISBN: (纸本)9783031664304;9783031664311

Recently, due to the rapid development of generative AI technologies, the use of AI-generated images has increased significantly, making the distinction between real and fake images crucial. Generative images may be used in various ways such as data training and fast image generation, but a potential for misuse, such as in deep fake or spreading false information, still exists. This study explores a novel model using the architecture of Swin-Transformer to distinguish between fake and real images generated based on CNN (Convolutional neural networks) and GAN (Generative Adversarial networks). The Swin-Transformer, a successor model of Vision in Transformer (ViT), applies the structure of the Transformer, which has shown outstanding performance in natural language processing, to the field of images and demonstrates excellent pixel-level segmentation performance. Real and fake images require detailed pixel-level analysis, in which the Swin-Transformer exhibits higher accuracy. Improving the performance of distinguishing between real and fake images is expected to set limits on indiscreet image generation, bringing further effects such as preventing the indiscriminate use of AI images through program-based discrimination/legal sanctions.

关键词： artificial intelligence Convolution neural network Generative adversarial network Real and fake

来源：评论

学校读者我要写书评

暂无评论

The use of neural networks and quantum algorithms for the classification images from UAV 7

The use of neural networks and quantum algorithms for the cl...

引用

7th IEEE International conference on Actual Problems of Unmanned Aerial Vehicles Development, APUAVD 2024

作者： Roman, Pavlov National Technical University of Ukraine Igor Sikorsky Kyiv Polytechnic Institute Kyiv Ukraine

ISBN: (纸本)9798331534141

The work explores the use of artificial intelligence and quantum computing for UAV video processing. A method of image preprocessing, feature detection and subsequent application of convolutional networks is proposed. Experiments show that quantum-augmented neural networks outperform traditional methods in terms of accuracy and speed, especially in binary tasks, highlighting the potential of quantum computing for UAV image analysis. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

A learned pixel-by-pixel lossless image compression method with 59K parameters and parallel decoding

引用

MULTIMEDIA TOOLS AND applications 2023年第8期83卷 22975-22993页

作者： Guemues, Sinem Kamisli, Fatih Middle East Tech Univ Elect & Elect Engn TR-06800 Ankara Turkiye

This paper considers lossless image compression and presents a learned compression system that can achieve state-of-the-art lossless compression performance but uses only 59K parameters, which is one or two order of magnitudes less than other learned systems proposed recently in the literature. The explored system is based on a learned pixel-by-pixel lossless image compression method, where each pixel's probability distribution parameters are obtained by processing the pixel's causal neighborhood (i.e. previously encoded/decoded pixels) with a simple neural network comprising 59K parameters. This causality causes the decoder to operate sequentially, i.e. the neural network has to be evaluated for each pixel sequentially, which increases decoding time significantly with common GPU software and hardware. To reduce the decoding time, parallel decoding algorithms are proposed and implemented. The obtained lossless image compression system is compared to traditional and learned systems in the literature in terms of compression performance, encoding-decoding times and computational complexity.

关键词： image compression artificial neural networks Entropy coding Gaussian mixture model

来源：评论

学校读者我要写书评

暂无评论

A fast method for load detection and classification using texture image classification in intelligent transportation systems

引用

MULTIMEDIA TOOLS AND applications 2024年第32期83卷 78609页

作者： Eghbal, Najmeh Anaraki, Behzad Ghayoumi Cheraghi-Shami, Farideh Sadjad Univ Dept Elect & Biomed Engn Mashhad Iran Islamic Azad Univ Dept Biomed Engn Mashhad Branch Mashhad Iran Roshan Tolou Shargh Co Mashhad Iran

The surveillance and management of cargo fleets is a crucial objective of intelligent transportation systems. Load, especially overload, has a destructive effect on roads and bridges, and monitoring it can increase the life of road surface and its structure. For low-end hardware with lack of CPU power and no GPU support, this paper presents a rapid method to detect whether heavy vehicles have loads or not;then it proposes a fast method for classifying load types to distinguish soil and construction waste from other miscellaneous loads for heavy weight vehicles. This paper applies a method for classifying cargo types using image processing and texture image classification. This method extracts features for statistical analysis of texture images based on gray-level co-occurrence matrices and local binary patterns. The classification is carried out by support vector machine, k-nearest neighborhood, K-mean, artificial neural networks and random forest classifiers. A large number of positive and negative patterns have been used to train these classifiers. We compare the performance of proposed extracted features and classifiers. The simulation results demonstrate that soil and construction waste can be identified from other miscellaneous loads effective in real-time implementation.

关键词： Intelligent transportation system Texture classification Machine learning Overload detection Real-time image processing

来源：评论

学校读者我要写书评

暂无评论

Augmented GAN: Advancing Generative Adversarial networks Through Enhanced Discriminator Training and Generator Stability

Augmented GAN: Advancing Generative Adversarial Networks Thr...

引用

2024 International conference on Signal processing, Computation, Electronics, Power and Telecommunication, IConSCEPT 2024

作者： Prakash, Aman Varghese, Ryan Binu Tanwar, Lavi Delhi Technological University Department of Electronics and Communications Engineering India

ISBN: (纸本)9798331540685

General Adversarial networks (GANs) have emerged as a powerful framework for generating reliable and transformative synthetic data in areas such as image generation, image and text synthesis, and data augmentation. This paper presents a comprehensive guide to building a Generative Adversarial neural Network using TensorFlow and Python. Through our research on this topic, we delve into the implementation details, providing step-by-step instructions for constructing both the generator and discriminator networks using TensorFlow, a popular deep-learning framework. Furthermore, we explore techniques for optimizing GAN performance, including architectural modifications, loss function selection, and training strategies. We demonstrate how to train a GAN on real-world datasets through practical examples, showcasing its capability to generate high-quality synthetic samples that closely resemble the training data distribution. In this paper, we aim to match fashion products such as clothing and accessories using GAN. This paper also focuses on improving the robustness and generalization of Convolutional neural Network (CNN) models, such as image classification and product recommendation. By combining image results with GANs and incorporating these artificial data into the training process, we aim to improve the performance of CNN models in real-world applications. © 2024 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Research on Small Target Detection Algorithm of Convolutional neural networks Based on Receptive Field 6

Research on Small Target Detection Algorithm of Convolutiona...

引用

6th International conference on artificial Intelligence and Computer applications

作者： Xue, Hao Southeast Univ Nanjing Peoples R China

ISBN: (纸本)9798350377255;9798350377262

With the continuous renovation of computer vision and deep learning technologies, object detection has emerged as a crucial research orientation in the contemporary field of artificial intelligence. Among numerous detection targets, small objects pose a technical challenge for detection due to their small pixel proportion, scarce information content, and vulnerability to external interferences and occlusions. This paper conducts an in-depth exploration of the key technologies for small object detection and proposes a small object detection algorithm model of convolutional neural network based on receptive fields. Through the improvement of the Inception module and the design of a novel receptive field enhancement structure, the convolutional neural network can cover the small object areas in the image more effectively during the feature extraction process, thereby enhancing the recognition ability for small objects. Simultaneously, a feature fusion module is established. The core mechanism of this module lies in adopting a multi-scale approach to transform and integrate the feature points extracted from each convolutional layer. By fusing the feature data of various network layers, this model can extract the details of the image more profoundly, thereby enhancing the detection accuracy of small objects. Compared with other similar methods in VOC 2007 test set, the proposed model has increased the mAP value by 0.5% to 18.4%, while maintaining a faster detection rate. On the MS COCO dataset, compared with other algorithms, the proposed algorithm has achieved an improvement of 1.6%similar to 13.2% in mAP value, which fully verifies the effectiveness and practical value of the algorithm.

关键词： Small target detection Convolutional neural network Feature fusion Receptive field Deep learning

来源：评论

学校读者我要写书评

暂无评论

L-MFFN: Lightweight Multiscale Feature Fusion Network with Limited Samples for HSI Classification

L-MFFN: Lightweight Multiscale Feature Fusion Network with L...

引用

9th International conference on Signal and image processing (ICSIP)

作者： Ali, Aamir Mu, Caihong Wang, Yafeng Liu, Yi Xidian Univ Sch Artificial Intelligence Xian 710071 Peoples R China Xidian Univ Sch Elect Engn Xian 710071 Peoples R China

ISBN: (纸本)9798350350920

Hyperspectral image (HSI) classification is valuable in remote sensing due to its rich spectral and spatial information. In the last decade, deep learning methods, especially Convolutional neural networks (CNNs), have revolutionized HSI classification by extracting intangible semantic features and maintaining the spatial structure during feature extraction. However, the efficacy of these techniques can be constrained by the limited availability of labeled samples in HSI data. To address the issue of small-sample HSI classification, a Lightweight Multiscale Feature Fusion Network (L-MFFN) is introduced. The Multiscale Feature Extraction Module (MFEM) and the enhanced Spectral-Spatial Attention Module (SSAM) are designed and combined in L-MFFN, optimizing the use of deep and shallow features. This integration improves the extraction and fusion of multiscale spectral-spatial features, enhancing classification performance. The proposed model demonstrates state-of-the-art performance across two HSI datasets and stands out in situations with limited labeled samples, highlighting its capability to effectively tackle the challenge of small-sample HSI classification.

关键词： hyperspectral image classification small sample multiscale feature fusion spectral-spatial attention

来源：评论

学校读者我要写书评

暂无评论

Applying Deep neural networks and NLP Techniques for Sentiment Analysis in Social Media Data 2

Applying Deep Neural Networks and NLP Techniques for Sentime...

引用

2nd International conference on artificial Intelligence and Machine Learning applications, AIMLA 2024

作者： Brinda, B.M. Rajan, C. Geetha, K. Nathiya, N. Raguraman, P.J. Srinivasan, K. Paavai College of Engineering Computer Science and Engineering Namakkal India Information Technology KSRCT College of Technology Namakkal India Excel Engineering College Computer Science and Engineering Namakkal India Paavai College of Engineering Artificial Intelligence and Data Science Namakkal India

ISBN: (纸本)9798350349221

This work offers a comprehensive investigation of sentiment analysis in social media communication through the integration of deep learning techniques with a natural language processing (NLP) methodology. The goal of the project is to create a matching model that can be used in real-world social processes. This model will allow for the precise identification of pertinent content that is dynamically changing and the realtime selection of key phrases based on available data. The paper builds a message sentiment analysis model and an image message multimodal sentiment analysis model, exploring unimodal and multimodal sentiment analysis algorithms in social networks. The study shows how well it works to extract complex emotions from social media writing by fusing advanced deep neural networks - like transformers or recurrent neural networks - with natural language processing techniques. Furthermore, the study presents noteworthy results, such as 96.1% order accuracy for brief texts in deep learning models that have been optimized, sentiment filtering for positive, neutral, and negative comments in social network data that has been successful, and an assessment of sophisticated semantic similarity models that offers a thorough comprehension of their performance in classification tasks. This work provides insightful information that may be used to improve sentiment analysis models and their usefulness in tasks involving semantic similarity and social network data interpretation. © 2024 IEEE.

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

LRGAN: Learnable Weighted Recurrent Generative Adversarial Network for End-to-End Shadow Generation

LRGAN: Learnable Weighted Recurrent Generative Adversarial N...

引用

International Joint conference on neural networks (IJCNN)

作者： Xue, Junsheng Huang, Hai Zhou, Zhong Xu, Shibiao Chen, Aoran Beijing Univ Posts & Telecommun Sch Informat & Commun Engn Beijing Peoples R China Beihang Univ State Key Lab Virtual Real Technol & Syst Beijing Peoples R China Beijing Univ Posts & Telecommun Sch Artificial Intelligence Beijing Peoples R China

ISBN: (纸本)9798350359329;9798350359312

In augmented reality(AR) applications, it is a challenging task to generate virtual object shadows while maintaining the precision and consistency of virtual and real areas. To achieve the above target, we propose a learnable weighted recurrent generative adversarial network(LRGAN) for end-to-end shadow generation. Without any additional computational overhead, LRGAN only needs to analyze the background context to create a bridge between the target shadows and the background. Our model incorporates multiple progressive steps to recurrently compute the precise reference masks, based on which a fine-grained shadow generation module generates the shadows. A learnable weighted fusion module, which can normalize pixel values to deal with pixel overflow, fuses the generated shadows with the original image. In addition, we adopt the combined method of module training and the whole model training. Experimental results show that our proposed LRGAN not only improves the plausibility of shadow location and shape but also achieves color harmony in the shadow areas. In the absence of other prior knowledge or post-processing, it outperforms the State-of-the-Art end-to-end methods.

关键词： virtual shadow generation augmented reality generative adversarial network weighted fusion recurrent structure

来源：评论

学校读者我要写书评

暂无评论

Prediction of Pathogen Causing Rice Plant Disease and Recommendation using Enhanced Machine Learning Technique 3

Prediction of Pathogen Causing Rice Plant Disease and Recomm...

引用

3rd International conference on Applied artificial Intelligence and Computing, ICAAIC 2024

作者： Priya, J Suji Priya, M Hema Iyswarya, M. Kiruthika, K. Sona College of Technology Department of Master of Computer Applications Salem India

ISBN: (纸本)9798350375190

Plant diseases pose significant challenges to agricultural productivity, impacting both the quality and quantity of crop yields. Early detection and effective management of these diseases are essential for mitigating their detrimental effects on agricultural output. However, traditional methods of disease monitoring and diagnosis are often costly and require continuous expert intervention. To address these challenges, this study proposes a novel approach that combines image processing techniques and advanced machine learning methodologies for the accurate identification and classification of rice leaf diseases. By utilizing datasets comprising images of diseased rice leaves, environmental data, and historical disease reports, the proposed system utilizes Convolutional neural networks (CNNs) for image analysis and feature extraction. Subsequently, a Support Vector Machine (SVM) classifier is employed for disease detection and classification based on the extracted features. The integration of image processing and machine learning technologies offers a cost-effective and efficient solution for early disease detection and management in agriculture. By automating the process of disease identification, the proposed system facilitates timely interventions, thereby minimizing crop losses and promoting agricultural sustainability. © 2024 IEEE.

关键词： Convolutional neural networks

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：