检索结果-内蒙古大学图书馆

Spike-Event-Driven Deep Spiking neural Network With Temporal Encoding

IEEE SIGNAL processing LETTERS 2021年 28卷 484-488页

作者： Zhang, Zhixuan Liu, Qi Univ Elect Sci & Technol China Sch Comp Sci & Engn Chengdu 610054 Peoples R China Natl Univ Singapore Dept Elect & Comp Engn Singapore 117583 Singapore

Feature extractionplays an important role before pattern recognition takes place. The existing artificial neural networks (ANNs), however, ignoreto learn and represent temporal information, instead of only utilizing spatial information for recognition. Moreover, the substantial computational and energy costs resulted from the conventional ANN-based classifiers, limit their uses in mobile and embedded applications. In this work, we develop a sparse temporal encoding method which exploits both spatial and temporal information. On the basis of spike-timing-dependent plasticity and multi-scale structure, the resulting temporal feature representation integrates with a temporal spiking neural network (SNN) classifier to achieve high efficiency of parallel computing for feature extraction. Experimental evaluation on four benchmark datasets from image classification and speech recognition tasks show the proposed SNN model yielding state-of-the-art accuracy.

关键词： Neurons Encoding Feature extraction Computational modeling Task analysis image coding Biological neural networks image classification multi-scale speech recognition spike-event-driven spiking neural network

来源：评论

学校读者我要写书评

暂无评论

Animal Video Retrieval System using image Recognition and Relationships Between Concepts of Animal Families and Species 5

Animal Video Retrieval System using Image Recognition and Re...

引用

5th IEEE International conference on image processing applications and Systems (IPAS)

作者： Watanabe, Chinatsu Kaneko, Mayu Chandrasiri, Naiwala P. Kogakuin Univ Intelligent Media Proc Lab Tokyo Japan Mizuho Res & Technol Ltd Tokyo Japan

ISBN: (纸本)9781665462198

In recent years, video streaming services have become increasingly popular. In general, the search function in a video sharing service site evaluates the relevance of a search query to the title, tags, description, and so on given by the creator of the video. Then, the search results with the highest relevance are displayed. Therefore, if a title is given to a video that does not match its content, there is a possibility that a video with low relevance will be found. In this research, ( 1) we built a new system that retrieves animal videos that are relevant to its content using image recognition. (2) By describing the relationships between the concepts of animal families and species and incorporating them into the retrieval system, it is possible to retrieve animal videos by their family names. Adding retrieval by animal family name enabled us to find species that have not been learned. In this research, (3) we confirmed the usefulness of our video retrieval system using trained neural networks, GoogLeNet and ResNet50, as animal species classifiers.

关键词： Video Retrieval image Recognition Deep Learning Animal Video Relationships between concepts

来源：评论

学校读者我要写书评

暂无评论

Application of Machine Learning Algorithms in artificial Intelligence image Recognition Technology

Application of Machine Learning Algorithms in Artificial Int...

引用

Intelligent Systems and Computational networks (ICISCN), International conference on

作者： Qi Ouyang Siyang Dai Kexin Xie Zeyu Sun Yao Wang Sustainable Development of Architecture and Landscape Polytechnic of Milan Milan Italy Faculty of Engineering and Information Technology The University of Melbourne Victoria Australia College of Business and Public Management Wenzhou-Kean University Wenzhou China School of Environment and Science Griffith University Brisbane Australia Nanjing Stomatological Hospital Affiliated Hospital of Medical School Institute of Stomatology Nanjing University Nanjing China

ISBN: (数字)9798331529246

ISBN: (纸本)9798331529253

In smart city applications such as medical image analysis, autonomous driving, and security monitoring, image recognition faces challenges like complex backgrounds, low-quality images, and diverse targets, affecting accuracy and robustness. This study explores machine learning algorithms to enhance image recognition performance by addressing noise interference, intricate backgrounds, and feature extraction difficulties. It combines Convolutional neural networks (CNN) with transfer learning, starting with data preprocessing to reduce noise, using pre-trained CNN models to extract high-level features, and fine-tuning with a ResNet transfer learning strategy for specific tasks. Additionally, ensemble learning methods are employed to further improve model robustness and accuracy. Experimental results show that the ensemble model maintains around 85% accuracy even with high background complexity, and transfer learning achieves 90% accuracy when the sample size reaches 1000. These findings demonstrate that transfer and ensemble learning effectively enhance image recognition accuracy and resilience in complex environments.

关键词： image recognition Accuracy Machine learning algorithms Transfer learning Noise Interference Feature extraction Robustness Ensemble learning Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Precision in Dermatology: Combining U-Net and Quantum neural networks for Melanoma Diagnosis

Precision in Dermatology: Combining U-Net and Quantum Neural...

引用

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

作者： Maria Frasca Ilaria Cutica Gabriella Pravettoni Davide La Torre Department of Oncology and Hemato-Oncology University of Milan Milan Italy SKEMA Business School Université Côte d'Azur Sophia Antipolis France

ISBN: (数字)9798350369106

ISBN: (纸本)9798350369113

Melanoma is one of the most severe types of skin cancer, and early diagnosis is crucial to improving the chances of successful treatment. Deep learning models, in particular, have proven to be a highly promising application of artificial intelligence in helping dermatologists diagnose melanoma early. By using these models, dermatological images can be analyzed with greater precision, making it easier to identify suspicious lesions and differentiate between benign and malignant ones. This study shows that more accurate segmentation and classification of skin lesions can be achieved by combining models like U-Net with preprocessing methods such as Autoencoder. This can lead to better melanoma detection and treatment. Additionally, we employed a hybrid CNN-quantum neural network model for classification, which achieved an accuracy of 99.67%, a precision of 99.35%, and a recall of 99.67%.

关键词： Deep learning image segmentation Accuracy neural networks Dermatology Autoencoders Melanoma Skin Lesions artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

Proposed CNN Model for Myanmar Traffic Sign Recognition System

Proposed CNN Model for Myanmar Traffic Sign Recognition Syst...

引用

Computer applications (ICCA), IEEE conference on

作者： Kay Thinzar Phu M.S. Godwin Premi Sathyabama Institute of Science and Technology Chennai India Faculty of Computer Science Myanmar Institute of Information Technology Mandalay Myanmar Department of ECE Sathyabama Institute of Science and Technology Chennai India

ISBN: (数字)9798350353624

ISBN: (纸本)9798350353631

artificial intelligence has involved as main part to construct smart cities and applications for autonomous system. Traffic Sign recognition is one of the important applications that can be used by autonomous cars. Traffic Sign recognition provides the valuable information about road status to drive safety. Although traffic sign cannot be recognized accurately under low-illumination conditions and time of the day. This system proposes enhancement method based on the brightness value of the image to enhance the low-illumination images. The HSV thresholding method is used to identify traffic sing location from the input image and convolution neural network is proposed to recognized the traffic signs. The system produces good results under night time, cloudy, drizzle rain weather. Myanmar Traffic Sign of red, yellow and blue color traffic signs are used in this system.

关键词： Rain image recognition image color analysis Smart cities Roads Clouds neural networks

来源：评论

学校读者我要写书评

暂无评论

An Adaptive High-Performance Quantization Approach for Resource-Constrained CNN Inference 4

An Adaptive High-Performance Quantization Approach for Resou...

引用

IEEE International conference on artificial Intelligence Circuits and Systems (AICAS) - Intelligent Technology in the Post-Pandemic Era

作者： Chin, Hsu-Hsun Tsay, Ren-Song Wu, Hsin-, I Natl Tsing Hua Univ Dept Comp Sci Hsinchu Taiwan

ISBN: (数字)9781665409964

ISBN: (纸本)9781665409964

Recent convolutional neural network (CNN) development continues to advance the state-of-the-art model accuracy for various applications. However, the enhanced accuracy comes at the cost of substantial memory bandwidth and storage requirements and demanding computational resources. Although in the past the quantization methods have effectively reduced the deployment cost for edge devices, it suffers from significant information loss when processing the biased activations of contemporary CNNs. In this paper, we hence introduce an adaptive high-performance quantization method to resolve the issue of biased activation by dynamically adjusting the scaling and shifting factors based on the task loss. Our proposed method has been extensively evaluated on image classification models (ResNet-18/34/50, MobileNet-V2, EfficientNet-B0) with imageNet dataset and object detection model (YOLO-V4) with COCO dataset. The results show that our 4-bit integer (INT4) quantization models achieve better accuracy than the state-of-the-art 4-bit models, and in some cases, even surpass the golden full-precision models. The final designs have been successfully deployed onto extremely resource-constrained edge devices for many practical applications.

关键词： deep neural network low bit-width quantization

来源：评论

学校读者我要写书评

暂无评论

Rice Leaf Disease Diagnosis Using Dense EfficientNet Model 1

引用

23rd International conference on Intelligent Systems Design and applications, ISDA 2023

作者： Roopa Devi, E.M. Shanthakumari, R. Rajadevi, R. Anusuyaa Harini Lokesh Department of Information Technology Kongu Engineering College Perundurai India Department of Artificial Intelligence and Machine Learning Kongu Engineering College Perundurai India

ISBN: (数字)9783031648472

ISBN: (纸本)9783031648465

Global food security and agricultural sustainability are seriously threatened by the paddy leaf diseases that is rapidly spread. Four of the most common widespread disease that affect rice leaf are Hispa, brown spot, Leaf blast and as well as Healthy leaf are identified. Quickly and correctly identifying a specific disease in order to implement effective strategies for its control and treatment. After the dataset performed the necessary pre-processing, a variety of deep learning algorithms, comprising VGG19, Resnet50, and InceptionV3, were applied on it. Our deep learning model is based on the EfficientNetB3 architecture, which has demonstrated excellent performance in image analysis applications. The EfficientNetB3 model is trained on the collected dataset, employing transfer learning to leverage pre-trained weights and accelerate convergence. To enhance the discriminative power of the model, employ feature extraction techniques. These techniques extract relevant features from the images and provide valuable information for disease detection. The integration of deep learning and feature extraction techniques, specifically using EfficientNetB3 demonstrates the potential of modern technologies in the field of agriculture. The ultimate objective of this integrated technique is to enable agricultural precision by providing farmers with a technology that is capable of identifying and monitoring plant diseases precisely and efficiently. It aids to sustainable agricultural production by providing focused interventions and minimising resource use. In summary, the originality in this work is found in the careful coupling of cutting-edge deep learning architecture, transfer learning, and feature extraction approaches, showcasing the advances being made in disease identification for the improvement of agricultural sustainability. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

SEPLN-PD 2022 - Proceedings of the Annual conference of the Spanish Association for Natural Language processing: Projects and Demonstrations, co-located with the conference of the Spanish Society for Natural Language processing, SEPLN 2022

SEPLN-PD 2022 - Proceedings of the Annual Conference of the ...

引用

2022 Annual conference of the Spanish Association for Natural Language processing: Projects and Demonstrations, SEPLN-PD 2022

The proceedings contain 21 papers. The topics discussed include: exploring the generation of online content by users and its predictive influence on the relational quality. application to the Andalusian hotel sector;LIVING-LANG: living digital entities by human language technologies;ESAN: automating medical scribing in Spanish;big hug: artificial intelligence for the protection of digital societies;an exploration of the semantic knowledge in vector models: polysemy, synonymy and idiomaticity;ALIADA: artificial intelligence-based language applications for the detection of aggressiveness in social networks;exploring gender bias in Spanish deep learning models;and a neural machine translation system for Galician from transliterated Portuguese text.

关键词：

来源：评论

学校读者我要写书评

暂无评论

PETNet–Coincident Particle Event Detection using Spiking neural networks

PETNet–Coincident Particle Event Detection using Spiking Ne...

引用

Neuro Inspired Computational Elements conference (NICE)

作者： Jan Debus Charlotte Debus Günther Dissertori Markus Götz Institute for Particle Physics and Astrophysics ETH Zürich Switzerland Scientific Computing Center Karlsruhe Institute of Technology Germany Helmholtz AI

ISBN: (数字)9798350390582

ISBN: (纸本)9798350390599

Spiking neural networks (SNN) hold the promise of being a more biologically plausible, low-energy alternative to conventional artificial neural networks. Their time-variant nature makes them particularly suitable for processing time-resolved, sparse binary data. In this paper, we investigate the potential of leveraging SNNs for the detection of photon coincidences in positron emission tomography (PET) data. PET is a medical imaging technique based on injecting a patient with a radioactive tracer and detecting the emitted photons. One central post-processing task for inferring an image of the tracer distribution is the filtering of invalid hits occurring due to e.g. absorption or scattering processes. Our approach, coined PETNet, interprets the detector hits as a binary-valued spike train and learns to identify photon coincidence pairs in a supervised manner. We introduce a dedicated multi-objective loss function and demonstrate the effects of explicitly modeling the detector geometry on simulation data for two use-cases. Our results show that PETNet can outperform the state-of-the-art classical algorithm with a maximal coincidence detection F 1 of 95.2%. At the same time, PETNet is able to predict photon coincidences up to 36 times faster than the classical approach, highlighting the great potential of SNNs in particle physics applications.

关键词： Geometry High energy physics Tensors Computational modeling Detectors Prediction algorithms Data models

来源：评论

学校读者我要写书评

暂无评论

Using Convolutional neural networks for Detecting Acrylamide in Biscuit Manufacturing Process

Using Convolutional Neural Networks for Detecting Acrylamide...

引用

11th International conference on Data Science, Technology and applications (DATA)

作者： Topcuoglu, Dilruba Mentes, Berat Utkan Askin, Nur Sengul, Ayse Damla Cankut, Zeynep Deniz Akdemir, Talip Ayvaz, Murat Kurt, Elif Erdohan, Ozge Temiz, Tumay Ceylan, Murat Most Technol Yildiz Holding Adv Analyt Solut Istanbul Turkey Yildiz Holding Northstar Innovat Istanbul Turkey Konya Tech Univ Dept Elect Elect Engn Konya Turkey

ISBN: (纸本)9789897585838

Based on a research in 2002 (Ozkaynak & Ova, 2006), acrylamide substance is formed when excessive heat treatment (e.g. frying, grilling, baking) is applied to starch-containing products. This substance contains carcinogenic and neurotoxicological risks for human health. The acrylamide levels are controlled by random laboratory sampling. This control processes which are executed by humans, cause a prolonged and error prone process. In this study, we offer a Convolutional neural Network (CNN) model, which provides acceptable precision and recall rates for detecting acrylamide in biscuit manufacturing process.

关键词： Acrylamide Deep Learning image processing CNN Algorithm

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：