检索结果-内蒙古大学图书馆

Canadian conference on Electrical and Computer Engineering (CCECE)

作者： Firas Al-Ali Thilina Doremure Gamage Hewa WTS Nanayakkara Farhad Mehdipour Sayan Kumar Ray Manukau Institute of Technology Auckland New Zealand Otago Polytechnic AIC Auckland New Zealand

ISBN: (数字)9781728154428

ISBN: (纸本)9781728154435

Convolutional neural networks (CNNs) require massive parallelism due to the high-precision floating-point arithmetic operations they perform. So, demand of processing power in them is significantly higher than what a standard CPU can offer. This has traditionally made CNNs more suited for running on a Graphics processing Unit (GPU). However, GPUs consume much more power than CPUs, rendering the former impractical for implementing CNNs in Edge AI (artificial Intelligence), where restraining power consumption is paramount. On the other hand, FPGAs (Field Programmable Gate Arrays) are more suited for AI computing at the edge as they consume much lesser power than GPUs and even CPUs. Additionally, GPUs and CPUs are not suited for real-time AI applications, which require both high throughput and low latency at the same time and FPGAs excel at all these requirements. The purpose of this paper is to provide a study of the performance of FPGA as the most suitable platform for AI-based computing at the edge. To achieve this, we chose AlexNet, a popular CNN image classifier, for which we present a case study on four different platforms: CPU, GPU, embedded RISC core, and FPGA fabric. Then, we quantitatively measure and compare the performance in terms of inference time (time needed to classify an image) on all these platforms. Inference time using FPGA is reduced by almost 64, 1.6, and 1.1 times compared to a dual-core ARM, an i5-6400 CPU, and an Nvidia GPU, respectively.

关键词： Field programmable gate arrays Graphics processing units artificial intelligence Fabrics Python Hardware Central processing Unit

来源：评论

学校读者我要写书评

暂无评论

AI Dermatologist: Skin Cancer Detection with Deep Learning Technique

AI Dermatologist: Skin Cancer Detection with Deep Learning T...

引用

International conference on Computing and Networking Technology (ICCNT)

作者： Afraz Ul Haque Rupak Sadman Sadik Khan Alaya Parvin Alo Rupa Ghosh Md Fatin Ishrak Md Sanowar Hossain Sabuj Department of Computer Science and Engineering Daffodil International University Dhaka Bangladesh Department of Computer Science and Engineering Khulna University Khulna Bangladesh Department of Electrical and Electronic Engineering North South University Dhaka Bangladesh College of Business Administration International University of Business Agriculture and Technology Dhaka Bangladesh

ISBN: (数字)9798350370249

ISBN: (纸本)9798350370270

This paper explores the application of deep learning technologies in the detection and diagnosis of skin cancer, leveraging advanced convolutional neural networks (CNNs) such as VGG-16, ResNet50, and MobileNetV2. Given the increasing incidence of skin cancer worldwide, there is a critical need for more accurate and efficient diagnostic methods. Our study employs transfer learning to enhance the capability of these networks, enabling them to detect skin cancer from dermoscopic images with a high degree of accuracy, often surpassing human expert performance. We detail the process of data collection, preprocessing, model training, and evaluation, emphasizing the adaptation of these models to skin cancer detection. The results demonstrate significant improvements in diagnostic accuracy, with MobileNetV2 showing notable efficiency in processing and resource utilization—qualities that are ideal for real-time diagnostic applications. In MobileNetV2 we achieved 97% accuracy. Furthermore, the study addresses the challenges of adversarial attacks and the integration of non-image data to refine diagnostic procedures. Our findings highlight the transformative potential of AI in medical diagnostics, suggesting that deep learning could become a cornerstone in the early detection and treatment of skin cancer, thereby improving patient outcomes substantially.

关键词： Deep learning Training Adaptation models Accuracy Transfer learning Real-time systems Convolutional neural networks artificial intelligence Skin cancer Residual neural networks

来源：评论

学校读者我要写书评

暂无评论

Two-stage Train Components Defect Detection Based on Prior Knowledge

Two-stage Train Components Defect Detection Based on Prior K...

引用

IEEE/RSJ International conference on Intelligent Robots and Systems (IROS)

作者： Gang Peng Zhiyong Li Shaowei Wan Zhang Deng Key Laboratory of Image Processing and Intelligent Control Ministry of Education School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China Wuhan Lisai Technology Co. Ltd. Wuhan China

The existing method of detecting defects in train components, which relies on visual identification, requires extensive involvement from inspectors and presents certain limitations. In this study, a two-stage defect detection based on prior knowledge was developed, which first detects the types and positions of components, and then conducts targeted detection of possible existing defect types. The algorithm introduces the prior knowledge of the relative spatial position relationship of components and optimizes the detection of sub-components by cascaded convolutional neural networks and local scale-up. In this study, three methods were used, including deep learning, template matching, and quantitative evaluation based on prior knowledge, to perform targeted detection of defect types that may occur in components. Experiments have verified the adaptability and accuracy of the method, demonstrating its high value for engineering applications.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Human Action Recognition using Pre-trained Convolutional neural networks 20

Human Action Recognition using Pre-trained Convolutional Neu...

引用

2020 2nd International conference on Video, Signal and image processing

作者： Cherry Aly Fazly Salleh Abas Hock Ann Goh Multimedia University Malaysia

ISBN: (纸本)9781450388931

Recognition of human action is one of the challenges in the field of artificial intelligence. Deep learning model has become a research issue in action recognition applications due to its ability to outperform traditional machine learning approaches. The Convolutional neural Network is one of the architectures commonly used in most action recognition works. There are different models in the Convolutional neural Network, but no study has been done to evaluate which model has the best performance in understanding human actions. Thus, in this paper, we compare the performance of two separate pre-trained models of deep Convolutional neural Network in classifying the human actions to identify the different behaviours. GoogleNet and AlexNet are the used two models with fine-tuned parameters used for comparison, in addition, to use Long-Short Term Memory for the video's labels prediction. The paper's main contribution is that it offers a performance analysis of two separate fine-tuned deep CNN pre-trained models compared to the results of other recently proposed human action recognition methods applied on KTH, Weizmann, UCF11(YouTube actions) and UCF-Sports datasets.

关键词： Action recognition AlexNet Convolutional neural network (CNN) Deep learning GoogleNet Long-Short Term Memory (LSTM)

来源：评论

学校读者我要写书评

暂无评论

DESIGN & IMPLEMENTATION OF ACCELERATORS FOR APPLICATION SPECIFIC neural processing

arXiv

引用

arXiv 2024年

作者： Mayannavar, Shilpa Wali, Uday V. KLE DR. MSS CET Belagavi India Visvesvaraya Technological University Karnataka Belagavi India

Internet of Things (IoT) and artificial Intelligence (AI) are quickly changing the computing eco system. Evolution of IoT has been slow, possibly due to lack of fine-grain standardization. On the other hand, developments in AI have been accelerating with a pace unprecedented in history of computing. Modern AI is largely driven by new structures in artificial neural networks (ANN) like Convolutional neural Network (CNN) and Long Short Term Memory (LSTM). Considering the types of neurons in mammalian Central Nervous System (CNS), it is easy to see that many new ANN structures need to evolve to attain the ultimate goal of artificial General Intelligence (AGI). It is necessary to implement massively parallel architectures to realize modern AI systems, which in turn forces a complete overhaul of computing technology;everything ranging from processor design to end applications need to be reinvented. Primary motivation for this work was the need to implement hardware accelerators for a newly proposed ANN structure called Auto Resonance Network (ARN) for robotic motion planning. ARN is an approximating feed-forward hierarchical and explainable network. It can be used in various AI applications but the application base was small. Therefore, the objective of the research was twofold: to develop a new application using ARN and to implement a hardware accelerator for ARN. As per the suggestions given by the Doctoral Committee, an image recognition system using ARN has been implemented. An accuracy of around 94% was achieved with only 2 layers of ARN. The network also required a small training data set of about 500 images. Publicly available MNIST dataset was used for this experiment. All the coding was done in Python. Details and results of this work have been reported in an international conference in April 2019 and submitted to a journal. During the present work, sigmoid based resonator has been generalized to improve controllability of the resonator. A relation between cover

关键词： Robot programming

来源：评论

学校读者我要写书评

暂无评论

Application of Machine Learning and Rule Induction in Various Sectors

Application of Machine Learning and Rule Induction in Variou...

引用

Decision Aid Sciences and Application (DASA), International conference on

作者： Bashar Yaser Almansour Ammar Yaser Almansour Jamshaid Iqbal Janjua Muhammad Zahid Tahir Abbas Faculty of Business and Management Muscat University Muscat Oman Department of Administrative and Financial Sciences Oman College of Management and Technology Barka Oman Al-Khawarizimi Institute of Computer Science (KICS) University of Engineering & Technology (UET) Lahore Pakistan Department of Computer Science The TIMES Institute Multan Pakistan

ISBN: (数字)9798350369106

ISBN: (纸本)9798350369113

The far-reaching impact of machine learning, a key component of artificial intelligence, has transformed various sectors and facets of life through its extensive applications. This paper delves into the broad spectrum of machine learning applications, encompassing image and speech recognition, natural language processing, predictive analytics, personalized recommendations, and self-driving vehicles. Additionally, we investigate its influence on critical areas such as healthcare, finance, customer service, and cybersecurity. The paper also explores machine learning's potential to tackle intricate challenges like climate change and medical diagnosis, while shedding light on the future trajectory of this rapidly advancing field. By highlighting machine learning's adaptability and capabilities, this paper seeks to offer a thorough examination of its applications, fostering further innovation and investigation.

关键词： Analytical models Accuracy Biological system modeling neural networks Finance Machine learning Data models Trajectory Genetic algorithms Context modeling

来源：评论

学校读者我要写书评

暂无评论

Scalable Non-Cartesian Magnetic Resonance Imaging with R2D2

Scalable Non-Cartesian Magnetic Resonance Imaging with R2D2

引用

European Signal processing conference (EUSIPCO)

作者： Yiwei Chen Chao Tang Amir Aghabiglou Chung San Chu Yves Wiaux Institute of Sensors Signals and Systems Heriot-Watt University Edinburgh United Kingdom EPCC University of Edinburgh Edinburgh United Kingdom

ISBN: (数字)9789464593617

ISBN: (纸本)9798331519773

We propose a new approach for non-Cartesian magnetic resonance image reconstruction. While unrolled architectures provide robustness via data-consistency layers, embedding measurement operators in Deep neural Network (DNN) can become impractical at large scale. Alternative Plug-and-Play (PnP) approaches, where the denoising DNNs are blind to the measurement setting, are not affected by this limitation and have also proven effective, but their highly iterative nature also affects scalability. To address this scalability challenge, we leverage the “Residual-to-Residual DNN series for high-Dynamic range imaging (R2D2)” approach recently introduced in astronomical imaging. R2D2's reconstruction is formed as a series of residual images, iteratively estimated as outputs of DNNs taking the previous iteration's image estimate and associated data residual as inputs. The method can be interpreted as a learned version of the Matching Pursuit algorithm. We demonstrate R2D2 in simulation, considering radial k-space sampling acquisition sequences. Our preliminary results suggest that R2D2 achieves: (i) suboptimal performance compared to its unrolled incarnation R2D2-Net, which is however non-scalable due to the necessary embedding of NUFFT-based data-consistency layers; (ii) superior reconstruction quality to a scalable version of R2D2-Net embedding an FFT-based approximation for data consistency; (iii) superior reconstruction quality to PnP, while only requiring few iterations.

关键词： Training Three-dimensional displays Magnetic resonance imaging Scalability Matching pursuit algorithms artificial neural networks Signal processing Extraterrestrial measurements Robustness image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Progressive Gradient Pruning for Classification, Detection and Domain Adaptation

Progressive Gradient Pruning for Classification, Detection a...

引用

International conference on Pattern Recognition

作者： Le Thanh Nguyen-Meidine Eric Granger Madhu Kiran Marco Pedersoli Louis-Antoine Blais-Morin Laboratoire d'imagerie de vision et d'intelligence artificielle (LIVIA) Eacute cole de technologie supérieure Montreal Canada Genetec Inc. Montreal Canada

Although deep neural networks (NNs) have achieved state-of-the-art accuracy in many visual recognition tasks, the growing computational complexity and energy consumption of networks remains an issue, especially for applications on platforms with limited resources and requiring real-time processing. Filter pruning techniques have recently shown promising results for the compression and acceleration of convolutional NNs (CNNs). However, these techniques involve numerous steps and complex optimisations because some only prune after training CNNs, while others prune from scratch during training, by integrating sparsity constraints or by modifying the loss function. In this paper, we introduce a new Progressive Gradient Pruning (PGP) technique for iterative filter pruning during training. In contrast to previous progressive pruning techniques, it relies on a novel filter selection criterion that measures the change in filter weights, uses a new hard and soft pruning strategy, and effectively adapts momentum tensors during the backward propagation pass. Experimental results obtained after training various CNNs on benchmark datasets for image classification, object detection and domain adaptation indicate that our PGP technique can achieve a better trade-off between classification accuracy and network (time and memory) complexity than PSFP and other state-of-the-art filter pruning techniques. Code is available on GitHub link: https://***/Anon6627/Pruning-PGP.

关键词： Training Backpropagation Weight measurement Visualization Tensors Object detection artificial neural networks

来源：评论

学校读者我要写书评

暂无评论

Vision-Based Drug Classification System Using Transfer Learning Approach

Vision-Based Drug Classification System Using Transfer Learn...

引用

International conference on Business and Industrial Research (ICBIR)

作者： Patpon Laohakijvitoon Sivapong Nilwong Master of Engineering Program in Engineering Technology Thai-Nichi Institute of Technology Bangkok Thailand Department of Computer Engineering and Artificial Intelligence Thai-Nichi Institute of Technology Bangkok Thailand

ISBN: (数字)9798350383027

ISBN: (纸本)9798350383034

Dispensing generic drugs instead of original drugs from pharmacies may cause adverse drug events to patients. It is difficult to separate between generic drugs and originals due to their similarities, except for imprints on the tablets. In this paper, a vision-based medicine tablets classification system using Convolutional neural networks (CNN) with a transfer learning approach is proposed. The proposed classification system aims to helps its user distinguish between the original drugs and generic drugs based on imprints on the medicine tablets. Multiple CNN models are created and tested as the classifier for the drug classification system. Three well-known CNN models are chosen as base models for creating CNNs with the transfer learning, including VGG16, Inception-V3, and ResNet50-V2. The developed CNN models are tested with images of medicine tablet taken from smartphones. Results from the experiments are evaluated through accuracies, precisions, recalls, and F-1 scores. Experimental results show satisfying performances of the proposed system in classifying original drugs and generic drugs using visual information.

关键词： Drugs Visualization image processing Transfer learning Data preprocessing Medical services Mobile applications Convolutional neural networks Smart phones Biomedical imaging

来源：评论

学校读者我要写书评

暂无评论

The smart device for healthcare service: Iris diagnosis application

The smart device for healthcare service: Iris diagnosis appl...

引用

International conference on ICT and Knowledge Engineering

作者： David Habsara Hareva Samuel Lukas Natanael Oktavian Suharta Informatics Universitas Pelita Harapan Tangerang Indonesia

We have developed an iridology application for predicting a person health through analysis of iris image. Commonly for commercial purpose, it used Desktop-PC, but we embedded it on Android-based Smartphone. The objective of this research is to find out whether healthcare application that providing direct diagnosis from real-time capturing image can be carried out through mobile phone in general. For that purpose, we developed other version of application that run on windows mobile and windows operating system. The algorithm and structure of the application is made as closely as possible with a previous application. Those applications have been tested on different environment, such as operating system, programming language, and mobile device. The device tested with regard to product maker, processor speed, screen size, storage size, and image retrieval using built-in camera. Although the results can be predicted that the used device with high specifications will be faster to process image analysis, better to obtain images, clearly in displaying images, however finding out the low limit of requirement for running the application is essential. It is found using low camera specification to take the image of iris is not recommended. Additional macro lens causes difficulty when taking a picture. processing of iris image to predict health condition could be handled by 1 GHz processor, 512 internal memories, and 3.5 inches screen size, even it used low-end product.

关键词： Iris artificial neural networks Lungs image edge detection Mobile communication Liver Heart

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：