检索结果-内蒙古大学图书馆

38th AAAI Conference on Artificial Intelligence (AAAI) / 36th Conference on Innovative applications of Artificial Intelligence / 14th Symposium on Educational Advances in Artificial Intelligence

作者： Pei, Wenjie Xia, Tongqi Chen, Fanglin Li, Jinsong Tian, Jiandong Lu, Guangming Harbin Inst Technol Shenzhen Peoples R China Shenzhen Jiang Associates Creat Design Co Ltd Shenzhen Peoples R China Chinese Acad Sci Shenyang Inst Automat Shenyang Peoples R China

ISBN: (纸本)1577358872

As a prominent parameter-efficient fine-tuning technique in NLP, prompt tuning is being explored its potential in computer vision. Typical methods for visual prompt tuning follow the sequential modeling paradigm stemming from NLP, which represents an input image as a flattened sequence of token embeddings and then learns a set of unordered parameterized tokens prefixed to the sequence representation as the visual prompts for task adaptation of large vision models. While such sequential modeling paradigm of visual prompt has shown great promise, there are two potential limitations. First, the learned visual prompts cannot model the underlying spatial relations in the input image, which is crucial for image encoding. Second, since all prompt tokens play the same role of prompting for all image tokens without distinction, it lacks the fine-grained prompting capability, i. e., individual prompting for different image tokens. In this work, we propose the Spatially Aligned-and-Adapted visual Prompt model (SA2vP), which learns a two-dimensional prompt token map with equal (or scaled) size to the image token map, thereby being able to spatially align with the image map. Each prompt token is designated to prompt knowledge only for the spatially corresponding image tokens. As a result, our model can conduct individual prompting for different image tokens in a fine-grained manner. Moreover, benefiting from the capability of preserving the spatial structure by the learned prompt token map, our SA2vP is able to model the spatial relations in the input image, leading to more effective prompting. Extensive experiments on three challenging benchmarks for image classification demonstrate the superiority of our model over other state-of-the-art methods for visual prompt tuning. Code is available at https://***/tommy-xq/SA2vP.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

Rapid detection of imperfect wheat grains based on deep learning technique 2

Rapid detection of imperfect wheat grains based on deep lear...

引用

2nd International Conference on image processing, Computer vision and machine Learning, ICICML 2023

作者： Hu, Kui Zhang, Hongming Lyu, Bo Shen, Yongcai Fu, Jia Wang, Fudi Lin, Zichao Anhui University Institute of Physical Science and Information Technology Hefei China Chinese Academy of Sciences Institute of Plasma Physics HFIPS Hefei China University of Science and Technology of China Science Island Branch Graduate School Hefei China Hefei Normal University Hefei China

ISBN: (纸本)9798350331417

The detection and identification of imperfect wheat grains are of great significance in evaluating their quality. Manual inspection and separation of imperfect grains in wheat are time-consuming and expensive. Therefore, there is a need for a fast, automated, and accurate method to detect imperfect grains in wheat. In this study, we created an image acquisition platform to capture images of wheat grains. Each image was labeled as either spotted, sprouted, moldy, broken, or perfect grains. To balance calculation efficiency and prediction accuracy, it is necessary to identify a suitable model. Thus, we applied several mainstream deep-learning algorithms, including ResNet50, ResNet50 with an attention module, MobileNetv1, MobileNetv2, MobileNetv3-small, and MobileNetv3-large, to construct a model suitable for practical application. It was found that the MobileNetv3-Small model achieved a good balance between computational efficiency and prediction accuracy. The model based on the MobileNetv3-Small architecture achieved high accuracy, recall rate, and F1-score, all exceeding 96%. Although the MobileNetv3-Small model's accuracy is slightly lower than that of the ResNet50 model with an added attention mechanism, it significantly reduces computational costs, improves detection speed, and cuts prediction time by 50%. Compared with other models, the MobileNetv3-small-based model is more suitable for practical applications due to its advantages of high speed, high precision, and stable prediction performance. This study can provide technical guidance for intelligent recognition and detection of wheat grains in the future. © 2023 IEEE.

关键词： deep learning image pattern recognition MobileNetv3-small Wheat imperfect grains

来源：评论

学校读者我要写书评

暂无评论

Macro-Scale Pattern Recognition and Coordinate Identification in Real-time Spatio-temporal Overlap for Photonics Engineering applications

引用

IFAC-PapersOnLine 2024年第3期58卷 70-73页

作者： Haider Al-Juboori South East Technological University Faculty of Engineering Dept. of Electronics Engineering and Communications 806 Killeshin Building Kilkenny Road Carlow Ireland R93V960

The significance of high-speed machine vision in scientific and technological fields is growing, especially with the era of Industry 4.0 technologies. There are several pattern-matching algorithms that have various intriguing applications in ultralow-latency machine vision processing. However, the low frame rate of image sensors—which usually operate at tens of hertz—fundamentally limits the processing rate. The paper will conceptualize and develop the computerized pattern recognition technique that can be applied to investigate light beam profiles and extract the desired information according to the purpose required in this case study. In the current work, the automatic detection and inspection of laser spots were designed to perform analysis and alignment for the laser beam in comparison with the electron spot beam using the LabvIEW graphical programming environment, especially when the laser and electron beams overlap. This is one of the important steps for realizing the fundamental aim of test-FEL to produce short wavelengths with the second, third, and fifth harmonics at 131.5, 88, and 53 nm, respectively. The tentative version of the program achieved the elementary purpose, which fulfilled the accurate transversal alignment of the ultrashort laser pulses with the electron beam in the system of the FEL test facility at MAX-Lab, in addition to studying the beam’s stability and jittering range.

关键词： intelligent systems pattern matching real-time tracking computer vision concepts supporting control automation semi-robotic systems

来源：评论

学校读者我要写书评

暂无评论

machine Learning in Computer vision: A Review

引用

EAI ENDORSED TRANSACTIONS ON SCALABLE INFORMATION SYSTEMS 2021年第32期8卷 1-11页

作者： Khan, Abdullah Ayub Laghari, Asif Ali Awan, Shafique Ahmed Sindh Madressatul Islam Univ Fac Comp Sci Karachi Sindh Pakistan Benazir Bhutto Shaheed Univ Lyari Fac Comp Sci & Informat Technol Karachi Pakistan

INTRODUCTION: Due to the advancement in the field of Artificial Intelligence (AI), the ability to tackle entire problems of machine intelligence. Nowadays, machine learning (ML) is becoming a hot topic due to the direct training of machines with less interaction with a human. The scenario of manual feeding of the machine is changed in the modern era, it will learn automatically. Supervised and unsupervised ML techniques are used as a distinct purpose like feature extraction, pattern recognition, object detection, and classification. OBJECTIvES: In Computer vision (Cv), ML performs a significant role to extract crucial information from images. Cv successfully contributes to multiple domains, surveillance system, optical character recognition, robotics, suspect detection, and many more. The direction of Cv research is going toward healthcare realm, medical imaging (MI) is the emerging technology, play a vital role to enhance image quality and recognized critical features of binary medical image, covert original image into grayscale and set the threshold values for segmentation. CONTRIBUTION: This paper will address the importance of machine learning, state-of-the-art, and how ML is utilized in computer vision and image processing. This survey will provide details about the type of tools and applications, datasets, and techniques. Limitations of previous work and challenges of future work also discussed. Further, we identify and discuss a set of open issues yet to be addressed, for efficiently applying of ML in Computer vision and image process. METHODS, RESULTS, AND CONCLUSION: In this review paper, we have discussed the techniques and various types of supervised and unsupervised algorithms of ML, general overview of image processing and the results based on the impact;neural network enabled models, limitations, tools and application of Cv, moreover, highlight the critical open research areas of ML in Cv.

关键词： machine Learning Computer vision Supervised and Unsupervised Learning Medical Imaging Pattern Recognition Feature Extraction Neural Network

来源：评论

学校读者我要写书评

暂无评论

Maize Leaf Healthy and Unhealthy Classification Using image processing Technique and machine Learning Classifiers 4th

Maize Leaf Healthy and Unhealthy Classification Using Image ...

引用

4th International Conference on Communications and Cyber-Physical Engineering, ICCCE 2021

作者： Khade, vishnu C. Patil, Sanjay B. Jadhav, Sachin B. BSCOER Pune India AGCE Satara India SCSOCE Dhangawadi Pune India BVCOE Kolhapur India

ISBN: (纸本)9789811679841

Automatic detection of the healthy and unhealthy maize plant leaf is a prevalent machine vision learning task and has significant applications in the Food Industry. In this paper, effective machine learning technique for maize leaf healthy and unhealthy classifications based on leaf images that have been presented. This study estimates color feature extraction using RGB mean and standard deviation and the classification, using PNN and KNN methods. A new Five-stage image processing method is presented (including image pre-processing, image segmentation, feature extraction, classification, and performance analysis). The Experimental results show that a small set of RGB color features reach an accuracy of 92.5% and 90% using PNN and KNN classifier respectively, while doing classification the KNN classifier requires more computational time as compared to PNN Classifier. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： machine learning

来源：评论

学校读者我要写书评

暂无评论

A TinyML based Residual Binarized Neural Network for real-time image Classification 6

A TinyML based Residual Binarized Neural Network for real-ti...

引用

6th International Conference on Electronics, Communication and Aerospace Technology, ICECA 2022

作者： Srinivasan, C. Sridhar, P. Priya, v. Hari Swathi, S. Saveetha Engineering College Department of Ece Tamilnadu Chennai India Sri Ramakrishna Engineering College Department of Ece Tamilnadu Coimbatore India

ISBN: (纸本)9781665482714

image processing is an important requirement in IoT applications such as robotics, augmented reality, computer vision, Industry 4.0 etc. The capabilities of IoT devices for image processing are limited to sensing the environment, processing and communicating the results. Tiny machine learning (TinyML) is a new paradigm that takes advantage of the IoT to deploy deep learning models to perform complex tasks in resource constrained embedded devices. image classification is an important task in IoT to interpret images of a particular scene or class. Currently, this task is performed in embedded devices using Binarized Neural Networks (BNNs), which can be converted to a set of weights using a one-hot encoding process. These networks integrated with hardware accelerators can be trained to perform image processing tasks in real-time. This paper proposes a BNN for image classification based on residual learning paradigm, called Tiny-BNN which exploits the skip connections to reduce information loss, and improve the training time and accuracy. Experimental results show that the model achieves a classification accuracy of 90.1 % and 91.6% on the on CIFAR-10 and MNIST datasets respectively. © 2022 IEEE.

关键词： image classification

来源：评论

学校读者我要写书评

暂无评论

Evaluation of the Energy viability of Smart IoT Sensors Using TinyML for Computer vision applications: A Case Study 6th

Evaluation of the Energy Viability of Smart IoT Sensors U...

引用

6th IFIP International Conference on Internet of Things, IFIP IoT 2023

作者： De Nardi, Adriel Monti Monteiro, Maxwell Eduardo Instituto Federal do Espírito Santo Campus Serra ES Serra29166-630 Brazil

ISBN: (纸本)9783031458774

TinyML technology, situated at the intersection of machine Learning, Embedded Systems, and the Internet of Things (IoT), presents a promising solution for a wide range of IoT domains. However, achieving successful deployment of this technology on embedded devices necessitates optimizing energy efficiency. To validate the feasibility of TinyML on embedded devices, extensive field research and real-world experiments were conducted. Specifically, a TinyML computer vision model for people detection was implemented on an embedded system installed in a turnstile at a Federal Institute. The device accurately counts people, monitors battery levels, and transmits real-time data to the cloud. Encouraging results were obtained from the prototype, and experiments were performed using a lithium battery configuration with three batteries in series. Hourly voltage consumption analysis was conducted, and the findings were illustrated through graphical representations. The camera sensor prototype exhibited a consumption rate of 1.25 v per hour, whereas the prototype without the camera sensor displayed a more sustainable consumption rate of 0.93 v per hour. This field research contributes to advancing TinyML applications and enriching studies regarding its integration with IoT and computer vision. © 2024, IFIP International Federation for Information processing.

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

Comparing Binary Classification and Autoencoders for vision-Based Anomaly Detection in Material Flow 11

Comparing Binary Classification and Autoencoders for Vision-...

引用

11th CIRP Global Web Conference, CIRPe 2023

作者： Ruediger-Flore, Patrick Klar, Matthias Hussong, Marco Mukherjee, Avik Glatt, Moritz Aurich, Jan C. RPTU in Kaiserslautern P.O. Box 3049 Kaiserslautern67653 Germany

Increasing digitalization has revolutionized the manufacturing industry. Advances in sensor systems and artificial intelligence enable more efficient data collection and analysis. In this context, vision-based quality inspection for product quality assurance and cost reduction has become a critical part of manufacturing quality control processes. A relatively unexplored area in this context are Aanomalies in the material flow, which appear in various forms, from misplaced objects to positioning and rotation issues in all three dimensions. The complex nature of these features elaborates the use of advanced machine learning techniques. Conventional methods use binary classification to separate data into normal and anomalous classes. However, acquiring a sufficient amount of labeled anomaly data is still challenging. This work aims to investigate the effectiveness of binary classification and autoencoders for image-based anomaly detection in a frame-assembling material handling application. The two methods are evaluated in terms of their performance (ROC-AUC) in an anomaly detection task. Therefore, we designed a virtual test bench to create sufficient image data in a controlled setting. Here the rotation and position of the hook of the mounting device is varied to simulate normal and anomalous positionings. The study concludes that binary classification achieves higher accuracy than autoencoders. However, autoencoders require fewer data acquisition efforts and can detect subtle anomalies that binary classification may miss. In addition, autoencoders can improve over time with more data input but require more model engineering than image classification models. While the transferability of deep image classification models is shown in multiple applications, it remains an open question for autoencoders. The results suggest that the choice of which method to use depends on the specific requirements of the manufacturing process. © 2023 The Authors. Published by Elsevier B.v

关键词： image enhancement

来源：评论

学校读者我要写书评

暂无评论

Normalized Determinant Pooling Layer in CNNs for Multi-Label Classification 7

Normalized Determinant Pooling Layer in CNNs for Multi-Label...

引用

Computational Imaging vII 2023

作者： Giuliano, Alessandro Hilal, Waleed Alsadi, Naseem Yawney, John Gadsden, S. Andrew Lab HamiltonONL8S 4L8 Canada Adastra Corporation TorontoONM5J 2J2 Canada

ISBN: (数字)9781510661615

ISBN: (纸本)9781510661608

Convolutional neural networks (CNNs) are a widely researched neural network architecture that has demonstrated exemplary performance in image processing tasks and applications compared to other popular deep learning and machine learning methods resulting in state-of-the-art performance in many image processing tasks such as image classification and segmentation. CNNs operate on the principle of automated learning of filters or kernels in contrast with hand-crafted digital filters to extrapolate features from images effectively. This paper aims to investigate whether a matrix's determinant can be used to preserve information in CNN convolutional layers. Geometrically the absolute value of the determinant is defined as a scaling factor of the linear transformation resulting from matrix multiplication. When an image's size is reduced into a feature space through a convolutional layer of a CNN, some information is lost. The intuition is that the scaling factor that results from the determinant of the pooling layer matrix can enhance the feature space introducing scaling as a piece of information in the feature space as well as lost relations between adjacent pixels. © 2023 SPIE. All rights reserved.

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

AI based Plant Growth Monitoring System using Computer vision

AI based Plant Growth Monitoring System using Computer Visio...

引用

2023 IEEE Technology and Engineering Management Conference - Asia Pacific, TEMSCON-ASPAC 2023

作者： Bhamare, Archana Upadhyay, vivek Bansal, Payal Poornima University Dept of Electronics & Communication Engineering Jaipur India

ISBN: (纸本)9798350384659

This research work aims to develop an AI-based plant growth monitoring system using computer vision. By leveraging computer vision algorithms and artificial intelligence techniques, the system will enable real-time and automated assessment of plant growth and health. Traditional methods for plant growth monitoring are time-consuming and labor-intensive. The proposed system will address these limitations by analyzing images of plants using image processing algorithms and extracting relevant features such as leaf area, plant height and width. machine learning algorithms will be employed to train models capable of recognizing patterns and predicting plant growth based on the extracted features. The system will be designed to accommodate different plant types and environmental conditions. By providing accurate and timely insights into plant health and growth, this AI-based plant growth monitoring system will benefit farmers, researchers, and plant enthusiasts, enabling informed decision-making for plant care and optimizing growth conditions. This research work aims to revolutionize plant growth monitoring by offering a cost-effective, scalable, and efficient solution with wide applications in agriculture, horticulture and environmental research. © 2023 IEEE.

关键词： Computer vision

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：