检索结果-内蒙古大学图书馆

1st international conference on image processing machine learning and pattern recognition

作者： Zhang, Haoming Chinese Flight Test Estab Xian Peoples R China

ISBN: (纸本)9798400707032

Aircraft image classification has important application value in military, civil aviation and UAV surveillance. In recent years, YOLOv8 has performed well in many tasks as an advanced target detection model, but there is still room for improving its classification accuracy in specific tasks. In this paper, we enhance the YOLOv8 model for aircraft image classification by integrating the ConvNeXt module into its backbone network and incorporating the Coordinate Attention (CA) mechanism, thereby improving feature extraction and spatial perception capabilities. The ConvNeXt module significantly improves the efficiency of feature extraction through deep convolution operation and feature fusion strategy;The CA enhances the model's ability to capture spatial features by incorporating coordinate information effectively. In this paper, experiments are conducted on six types of aircraft image datasets, including in-flight UAVs, fighters, helicopters, missiles, airliners, and rockets. The experimental results show that the average classification accuracy of the improved model on the test set is improved by 3.5%. This indicates that the improvement strategy of introducing the ConvNeXt module and CA attention mechanism is effective in improving the classification performance of YOLOv8. The study in this paper not only verifies the effectiveness of the proposed method, but also provides new ideas and methods for future research on aircraft image classification.

关键词： YOLOv8 ConvNeXt CA attention aircraft image classification

来源：评论

学校读者我要写书评

暂无评论

Unsupervised learning of local mean gray values for image pre-processing 1st

Unsupervised learning of local mean gray values for image pr...

引用

1st international Workshop on machine learning and Data Mining in pattern recognition

作者： Jahn, H DLR Deutsch Zentrum Luft & Raumfahrt EV Inst Weltraumsensor & Planetenerkundung D-12489 Berlin Germany

ISBN: (纸本)3540665994

A parallel-sequential unsupervised learning method for image smoothing is presented which can be implemented with a Multi Layer Neural Network. In contrast to older work of the author which has used 4-connectivity of processing elements (neurons) leading to a very big number of recursions now each neuron of network lever t+1 is connected with (2M+1)*(2M+1) neurons of layer t guaranteeing a significant reduction of network layers with the same good smoothing results.

关键词： Neurons

来源：评论

学校读者我要写书评

暂无评论

A lightweight diffusion model for image generation based on improved MobileViT 24

A lightweight diffusion model for image generation based on ...

引用

1st international conference on image processing machine learning and pattern recognition

作者： Liu, Yifei Sun, Wenyun Nanjing Univ Informat Sci & Technol Sch Artificial Intelligence Nanjing Peoples R China

ISBN: (纸本)9798400707032

Aiming at the problem that the structural complexity of the backbone architecture of denoising diffusion probabilistic model (DDPM) leads to the inefficiency of the training of the model, a lightweight image generation method (MobileDiT) based on MobileViT is proposed, which takes MobileViT block as the backbone architecture of DDPM, and improves the computational efficiency while keeping the image quality by combining lightweight convolution with Transformer. The conditional mechanism for introducing conditional information into the model is also improved by replacing the traditional layer normalization in the Transformer block with adaptive layer normalization and initialising each block as a constant function, allowing the model to process conditional information more efficiently. The experiment results show that the proposed model reduces the FID-50K to 2.15 and improves the IS value compared with models such as style Generative Adversarial Network (styleGAN) and Ablative Diffusion Model (ADM). This shows that the proposed model can not only improve the computational efficiency, but also enhance the quality of the generated images.

关键词： image generation Denoising Diffusion Probabilistic Model (DDPM) Computational complexity Lightweight convolution Conditional mechanism

来源：评论

学校读者我要写书评

暂无评论

Water area extraction method based on Res2Net-Unet 24

Water area extraction method based on Res2Net-Unet

引用

1st international conference on image processing machine learning and pattern recognition

作者： Liu, Xiaoqun Li, Ning Commun Univ China Nanjing Peoples R China Vimicro Technology Corp Beijing Peoples R China

ISBN: (纸本)9798400707032

Regarding the issues that traditional water area extraction algorithms for remote sensing images require preset thresholds, are vulnerable to interference from mountain or vegetation shadows, and have poor robustness at different spatiotemporal scales, a model based on the Res2Net-UNet is proposed to extract water area from remote sensing images. The model adopts the UNet's encoder-decoder architecture, deepens the model's level with the ResNet network, and performs multi-scale feature extraction and fine-grained feature expression by combining Res2Net. Moreover, the model integrates an attention mechanism to increase the weight of key features and fully utilize the spatial and channel information in the images. After that, the model adds a skip connection to better fuse the encoder and decoder to preserve the detailed features of the image. Experimental findings on the GF-1 dataset demonstrate that the Res2Net-UNet model significantly enhances the accuracy of water extraction in remote sensing images.

关键词： water area extraction remote sensing image Res2Net-Unet attention mechanism

来源：评论

学校读者我要写书评

暂无评论

Face recognition with Kernel Principal Component Analysis and Support Vector machine 1

Face Recognition with Kernel Principal Component Analysis an...

引用

1st international conference on Informatics, Multimedia, Cyber and Information System (ICIMCIS)

作者： Liliana, Dewi Yanti Setiawan, I. Made Agus State Polytech Jakarta Dept Informat & Comp Depok Indonesia Udayana Univ Denpasar Indonesia Univ Pittsburgh Pittsburgh PA 15260 USA

ISBN: (纸本)9781728129303

machine learning and pattern recognition recently become a hot topic in computing world. This is due to the fast-growing of resources as well as techniques that make it easier to solve machine learning and pattern recognition problems. Problems that require machine learning solutions may be very simple for humans but actually can be very complex for machines to solve them. Face recognition is amongst those problems. Almost all human can easily recognize others without require specific knowledge to do it, different from machines which require its. This paper discussed face recognition task using machine learning strategies which involved Kernel Principal Component Analysis (KPCA) and Support Vector machine (SVM) to identify person. KPCA extracted features from 2D image input and produced the important features of an image input. The extracted face features are recognized by SVM by classifying human face according to their stored identity in a database. SVM, which was basically a binary classifier worked by using one-against-one strategy to compare the face feature vector of a single test image to the stored face image in a face image database. Experiment results on grayscale images with size 92x112 pixels gave 96.25% of accuracy rate. Hence, KPCA and SVM for face recognition is a robust machine learning method.

关键词： face recognition kernel eigenface kernel PCA machine learning pattern recognition support vector machine

来源：评论

学校读者我要写书评

暂无评论

Active face recognition with a hybrid approach

引用

pattern recognition LETTERS 1997年第9期18卷 933-946页

作者： Tistarelli, M Grosso, E Univ Genoa Dept Commun Comp & Syst Sci DIST Comp Vis Lab I-16145 Genoa Italy

The automatic detection of person's identity is a very interesting issue both in social and industrial environments. In this paper a system for automatic identity recognition from face images, is presented. The proposed approach is based on an hybrid iconic approach, where a first recognition score is obtained by matching a person's face against an eigen-space obtained from an image ensemble of known individuals. The hypothesis is then verified by computing the correlation of the gray level histogram of the new face image with the histograms of the subjects in the database,A selective attentional mechanism is applied to reduce the amount of information needed to describe a database of human faces. This is accomplished both at the task level, by performing planned fixations, and at the sensor level, by adopting a space-variant sampling of the images, By using a space-variant image geometry, the size of the database is considerably reduced and consequently also the processing time for recognition. (C) 1997 Elsevier Science B.V.

关键词： face recognition visual processing automated recognition active vision singular value decomposition

来源：评论

学校读者我要写书评

暂无评论

A 1 TOPS/W Analog Deep machine-learning Engine With Floating-Gate storage in 0.13 μm CMOS

引用

IEEE JOURNAL OF SOLID-stATE CIRCUITS 2015年第1期50卷 270-281页

作者： Lu, Junjie Young, steven Arel, Itamar Holleman, Jeremy Univ Tennessee Dept Elect Engn & Comp Sci Knoxville TN 37996 USA

An analog implementation of a deep machine-learning system for efficient feature extraction is presented in this work. It features online unsupervised trainability and non-volatile floating-gate analog storage. It utilizes a massively parallel reconfigurable current-mode analog architecture to realize efficient computation, and leverages algorithm-level feedback to provide robustness to circuit imperfections in analog signal processing. A 3-layer, 7-node analog deep machine-learning engine was fabricated in a 0.13 mu m standard CMOS process, occupying 0.36 mm(2) active area. At a processing speed of 8300 input vectors per second, it consumes 11.4 mu W from the 3 V supply, achieving 1x10(12) operation per second per Watt of peak energy efficiency. Measurement demonstrates real-time cluster analysis, and feature extraction for pattern recognition with 8-fold dimension reduction with an accuracy comparable to the floating-point software simulation baseline.

关键词： Analog signal processing current mode arithmetic deep machine learning floating gate neuromorphic engineering translinear circuits

来源：评论

学校读者我要写书评

暂无评论

Research on Task-based Requirements Modeling and Compositional Verification for Automated processing System 24

Research on Task-based Requirements Modeling and Composition...

引用

1st international conference on image processing machine learning and pattern recognition

作者： Li, Junhu Xu, Jiankun Xie, Kun Chen, Jie Ruan, Kun 29 Res Inst China Elect Technol Corp Beijing Peoples R China East China Normal Univ Beijing Peoples R China

ISBN: (纸本)9798400707032

Formal verification is playing an increasingly important role in the development of safety-critical systems, such as signal process systems. Through formal modeling and verification, the behavior and decision-making processes of safety-critical systems can be described more precisely, thereby ensuring the safety of the system. By utilizing formal models, comprehensive verification of these models can be conducted to identify and fix potential errors and defects in the early phase. However, as the scale of systems increases, the complexity of formal verification models correspondingly increases, potentially resulting in verification failures. To address this issue, this paper proposes a task-based requirements modeling and compositional verification approach for signal process systems. It reduces the complexity of formal verification models by only considering task related requirements. This approach not only effectively identifies task related errors but also ensures a certain level of verification efficiency. We also conduct a practical case study on an automated processing system to demonstrate the validity of our approach.

关键词： Automated processing systems Requirements modeling Task errors Compositional verification

来源：评论

学校读者我要写书评

暂无评论

Innovating Edge Detection with Self-Attention Mechanisms and Novel Post-processing Modules 24

Innovating Edge Detection with Self-Attention Mechanisms and...

引用

1st international conference on image processing machine learning and pattern recognition

作者： Zheng, Haotian East China Univ Sci & Technol Sch Informat Sci & Engn Shanghai Peoples R China

ISBN: (纸本)9798400707032

In previous studies on edge extraction, researchers have consistently sought to balance efficiency, quality, and stability. To address this, we propose an edge detection algorithm that leverages the self-attention mechanism and post-processing methods to enhance the quality and stability of edge extraction in image processing. In recent years, self-attention mechanisms have emerged as a core component in many cutting-edge models, particularly in the fields of Natural Language processing (NLP) and Computer Vision. They are capable of capturing long-range dependencies within the data. Consequently, we introduce the self-attention mechanism to enhance the capability of feature representation. By employing channel compression and scaling factors to stabilize the computation of attention scores, we ensure the reliability of the training process. Next, we combine Gaussian blurring with adaptive thresholding in the post-processing stage to reduce noise and detail interference, thereby improving edge detection accuracy. Experimental results on extensive benchmark datasets demonstrate that the proposed algorithm significantly outperforms existing methods in terms of edge detection accuracy and stability, particularly in the presence of complex backgrounds and noise. Additionally, the method exhibits strong computational efficiency, making it suitable for practical application. These advancements offer new insights and approaches for the development of edge detection technology.

关键词： Self-Attention Edge Detection Edge Extraction model

来源：评论

学校读者我要写书评

暂无评论

Effects and Comparison of different Data pre-processing techniques and ML and deep learning models for sentiment analysis: SVM, KNN, PCA with SVM and CNN 1

Effects and Comparison of different Data pre-processing tech...

引用

1st international conference on Artificial Intelligence Trends and pattern recognition, ICAITPR 2022

作者： Hafeez, Shoaib Kathirisetty, Nikhila Vardhaman College of Engineering Information Technology Hyderabad India

ISBN: (数字)9781665425216

ISBN: (纸本)9781665425216

In this paper, we have discussed different data pre-processing techniques and different machine learning and deep learning models which are used for sentiment analysis. The dataset used was 'Restaurant Reviews' We have compared the results of different results of SVM, KNN, PCA with SVM and CNN models. Each of the different pre-processed datasets was passed to different machine learning and deep learning models and the results were compared to find the most useful data pre-processing technique for a particular model, so we can save resources (time and money) by concentrating our resources on that particular data pre-processing technique for that model. © 2022 IEEE.

关键词： Support vector machines

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：