检索结果-内蒙古大学图书馆

An Ensemble learning Model for Detecting Soybean Seedling Emergence in UAV imagery

SENSORS 2023年第15期23卷 6662页

作者： Zhang, Bo Zhao, Dehao Heilongjiang Bayi Agr Univ Coll Informat & Elect Engn Daqing 163319 Peoples R China

Efficient detection and evaluation of soybean seedling emergence is an important measure for making field management decisions. However, there are many indicators related to emergence, and using multiple models to detect them separately makes data processing too slow to aid timely field management. In this study, we aimed to integrate several deep learning and image processing methods to build a model to evaluate multiple soybean seedling emergence information. An unmanned aerial vehicle (UAV) was used to acquire soybean seedling RGB images at emergence (VE), cotyledon (VC), and first node (V1) stages. The number of soybean seedlings that emerged was obtained by the seedling emergence detection module, and image datasets were constructed using the seedling automatic cutting module. The improved AlexNet was used as the backbone network of the growth stage discrimination module. The above modules were combined to calculate the emergence proportion in each stage and determine soybean seedlings emergence uniformity. The results show that the seedling emergence detection module was able to identify the number of soybean seedlings with an average accuracy of 99.92%, a R-2 of 0.9784, a RMSE of 6.07, and a MAE of 5.60. The improved AlexNet was more lightweight, training time was reduced, the average accuracy was 99.07%, and the average loss was 0.0355. The model was validated in the field, and the error between predicted and real emergence proportions was up to 0.0775 and down to 0.0060. It provides an effective ensemble learning model for the detection and evaluation of soybean seedling emergence, which can provide a theoretical basis for making decisions on soybean field management and precision operations and has the potential to evaluate other crops emergence information.

关键词： emergence evaluation unmanned aerial vehicle imagery ensemble learning model growth stages emergence proportion

来源：评论

学校读者我要写书评

暂无评论

Single image Dehazing for Efficient Search Exploration Using Machine learning Technique

Single Image Dehazing for Efficient Search Exploration Using...

引用

Intelligent Computing and Control for Engineering and Business Systems (ICCEBS)

作者： Ashwini A Angel Merlin Suji Banu Priya Prathaban Shamini G I Gadde Mukesh Narayana Department of ECE Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology Chennai India Department of CSE St. Xavier's Catholic College of Engineering Nagercoil India Department of Networking and Communication School of computing SRM institute of Science and Technology Chennai India Department of ECE Sathyabama Institute of Science and Technology Chennai India

image dehazing is frequently used as a preprocessing step to enhance image clarity in various computer vision tasks, such as object tracking, intelligent surveillance and face recognition. It helps in the restoration of hazy images. deep learning techniques have recently been utilized to significantly improve the quality of visual representation of images, but it takes a lengthy time to compute. In order to avoid the phenomenon of latency in the primary image analysis tasks, we have to monitor the time required for processing. An end-to-end modeling for real-time image dehazing network is proposed in this research paper. A complete network processing model for dehazing in real-time is proposed in this research paper. To record dependencies with reference to the positions and channels of the feature map, a non-local slice module is created. A region proposal network is also used to provide candidates with great sensitivity, hence raising true positive rates. The dark channel is estimated from the given input image along with the saliency of the detected image. The multi-scale feature maps are then used to suggest a false positive reduction module. An extensive comparative experiments were conducted using SANet's performance with a number of cutting-edge CNN-based detection techniques. The usefulness of our suggested method to gradually recover haze images utilising a coarse-to-fine strategy is demonstrated by evaluation results on SANet. The proposed method demonstrated real-time performance comparable to existing method in experimental findings using a publicly available dataset.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Exploring Support Vector Machine and K-Nearest Neighbors for Pigeonpea Leaf image Disease Detection and Classification

Exploring Support Vector Machine and K-Nearest Neighbors for...

引用

Innovation in Technology (ASIANCON), Asian Conference on

作者： G. G. Rajput Vanita Bhimappa Doddamani Dept. of Computer Science Karnataka State Akkamahadevi Women University Vijayapur Karnataka India

ISBN: (数字)9798350354218

ISBN: (纸本)9798350354225

Plant diseases create a significant threat to nearly all -important plant species. Pigeonpea (Cajanus cajan(L.) Millsp), a member of the Fabaceae family. In the semi-arid tropics and subtropics of Asia and Africa, it is a vital legume shrub. However, over 100 diseases, including bacteria, nematodes, viruses, and fungus, attack this crop and cause economic harm. Several times, farmers have been unable to recognize the real disease occurrence and have been unable to manage it, resulting in markable yield reduction with regard to both quality and quantity. Smart agriculture has been used to control pests and eradicate plant diseases. Majority of these methods rely on machine learning or deep learning models as a method of providing disease detection solutions. This paper presents an image processing and machine-learning approach to the discovery of Pigeonpea leaf disease and its classification. In order to test the model proposed, a comprehensive dataset with images of healthy and diseased Pigeonpea leaves was constructed. A unique fusion of CLAHE, Retinex, FFT, and Gaussian algorithm was used for the enhancement of Pigeonpea leaf image. Machine learning classifying techniques SVM and KNN were applied for leaf classification. Accuracy, F1-score, precision and recall were used to determine the efficacy of the proposed framework.

关键词： Support vector machines Smart agriculture Plant diseases Technological innovation Accuracy Fast Fourier transforms Biological system modeling Nearest neighbor methods real-time systems Testing

来源：评论

学校读者我要写书评

暂无评论

Research on Enhanced Perception and Attention-Driven Person Re-Identification 20

Research on Enhanced Perception and Attention-Driven Person ...

引用

20th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery, ICNC-FSKD 2024

作者： Wu, Cancan Li, Dong Zhang, Zhi College of Computer Science and Technology Wuhan University of Science and Technology Hubei Wuhan430065 China Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System Hubei Wuhan430065 China Big Data Science and Engineering Research Institute Wuhan University of Science and Technology Hubei Wuhan430065 China

ISBN: (纸本)9798350356328

When matching similarity among pedestrians in images, pedestrian re-identification algorithms are often disturbed by occlusions. A typical tactic is to improve the robustness of occlusion features in the model. However, this approach neglects the noise introduced by other pedestrians in the image, leading to limited learning of relevant information between images. To address these issues, this paper introduces a pedestrian re-identification model that leverages perception-enhanced and attention mechanisms. By enhancing the perception of non-pedestrian noise and erasing non-pedestrian features, we then employ a cross-attention mechanism to subtly embed global and local features for finer feature recognition. Finally, through context attention, we learn the interaction between pedestrian images to improve re-identification accuracy. Results from studies performed on the Market-1501, CUHK03-D, and Duke-ReID datasets indicate that the suggested model surpasses leading contemporary models regarding Rank-1 scores and mean Average Precision (mAP), while maintaining good robustness. © 2024 IEEE.

关键词： Reamers

来源：评论

学校读者我要写书评

暂无评论

FakeFaceDiscriminator: Discrimination of AI-Synthesized Fake Faces

FakeFaceDiscriminator: Discrimination of AI-Synthesized Fake...

引用

Security of Information and Networks (SIN), International Conference on

作者： Xufeng Liu Quoc Hoan Vu Priyanka Singh School of EECS The University of Queensland Brisbane Australia

ISBN: (数字)9798331509736

ISBN: (纸本)9798331509743

In recent years, the rapid improvement of deep learning technologies, particularly Generative Adversarial Net-works, has led to the proliferation of high-quality synthetic facial images and videos, commonly known as deepfakes. This study aims to evaluate and compare the performance of three prominent deep learning models - ResN et, EfficientNet, and Xception - in detecting synthetic faces. Using the deepfake Detection Challenge and FaceForensics++ datasets, we system-atically assess each model's capability to handle diverse and challenging scenarios, including blurred and dark images. Data augmentation techniques, such as random blurring, brightness adjustment, and contrast enhancement, were employed to im-prove the models' robustness. Additionally, we applied model- specific optimizations, including the integration of Squeeze-and- Excitation blocks in Res Net, compound scaling in EfficientNet, and multi-scale feature fusion in Xception. These enhancements significantly improved the models' accuracy and resilience against low-quality synthetic data. Our results indicate that EfficientNet and Xception outperform ResNet in both general and adverse conditions, with EfficientNet excelling in high-resolution image processing and Xception showing superior performance in fine- grained feature extraction. Furthermore, the introduction of pre- trained weights, multitask learning frameworks, and dynamic learning rate adjustments during training contributed to the models' enhanced performance.

关键词： deep learning deepfakes Analytical models Accuracy Data models real-time systems Security Face detection Faces Optimization

来源：评论

学校读者我要写书评

暂无评论

Multi-scale ResNet for real-time underwater object detection

引用

SIGNAL image AND VIDEO processing 2021年第5期15卷 941-949页

作者： Pan, Tien-Szu Huang, Huang-Chu Lee, Jen-Chun Chen, Chung-Hsien Natl Kaohsiung Univ Sci & Technol Dept Elect Engn Kaohsiung Taiwan Natl Kaohsiung Univ Sci & Technol Dept Telecommun Engn Kaohsiung Taiwan Met Ind Res & Dev Ctr MIRDC Kaohsiung Taiwan

An automatic underwater object recognition system is essential to reduce the costs of underwater inspection. In this study, we propose a novel convolutional neural network architecture that is trained on underwater video frames. This method is based on a modified residual neural network (ResNet) for underwater object detection. Multi-scale ResNet (M-ResNet), the modified method, improves efficiency by utilizing multi-scale operations for the accurate detection of objects of various sizes, especially small objects. The experimental results show that the proposed method yields an accuracy of 96.5% (mAP) in recognition performance. As a consequence, we propose a novel system for automatic object detection as an application for marine environments.

关键词： Marine object recognition deep learning Convolutional neural network Residual neural network

来源：评论

学校读者我要写书评

暂无评论

Q-Segment: Segmenting images In-Sensor for Vessel-Based Medical Diagnosis

Q-Segment: Segmenting Images In-Sensor for Vessel-Based Medi...

引用

IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS)

作者： Pietro Bonazzi Yawei Li Sizhen Bian Michele Magno ETH Zürich D-ITET Zürich Switzerland

ISBN: (数字)9798350383638

ISBN: (纸本)9798350383645

This paper addresses the growing interest in deploying deep learning models directly in-sensor. We present "Q-Segment", a quantized real-time segmentation algorithm, and conduct a comprehensive evaluation on a low-power edge vision platform with an in-sensors processor, the Sony IMX500. One of the main goals of the model is to achieve end-to-end image segmentation for vessel-based medical diagnosis. Deployed on the IMX500 platform, Q-Segment achieves ultra-low inference time in-sensor only 0.23 ms and power consumption of only 72mW. We compare the proposed network with state-of-the-art models, both float and quantized, demonstrating that the proposed solution outperforms existing networks on various platforms in computing efficiency, e.g., by a factor of 75x compared to ERFNet. The network employs an encoder-decoder structure with skip connections, and results in a binary accuracy of 97.25 % and an Area Under the Receiver Operating Characteristic Curve (AUC) of 96.97 % on the CHASE dataset. We also present a comparison of the IMX500 processing core with the Sony Spresense, a low-power multi-core ARM Cortex-M microcontroller, and a single-core ARM Cortex-M4 showing that it can achieve in-sensor processing with end-to-end low latency (17 ms) and power consumption (254mW). This research contributes valuable insights into edge-based image segmentation, laying the foundation for efficient algorithms tailored to low-power environments.

关键词： image segmentation Power demand Program processors image edge detection System performance Receivers real-time systems

来源：评论

学校读者我要写书评

暂无评论

An effectual underwater image enhancement using deep learning algorithm 5

An effectual underwater image enhancement using deep learnin...

引用

5th International Conference on Intelligent Computing and Control Systems, ICICCS 2021

作者： Ramkumar, G. Anitha, G. Suresh Kumar, M. Ayyadurai, M. Senthilkumar, C. Saveetha University Department of Electronics and Communication Engineering Chennai Tamil Nadu India

ISBN: (纸本)9781665412728

Digital image processing domain is growing day-by-day by introducing novel technologies to provide assistance for several applications such as robotic activities, underwater network formation, and so on. In particular, underwater image processing is considered as the crucial task in image processing industry due to the flow of light waves that are not in the specific and expected range under the water level. While image restoration technology can adequately consider removing this same haze from source images, they need to obtain several images from a certain place that prevent it from being used in a real-time system. To overcome this issue, a deep study approach is developed by providing excellent outcomes of deep learning approaches in several other image analysis concerns such as colorizing images or object identification. A convolution neural network (CNN) model is trained to de-haze the individual images with image restoration in order to perform further with an image improvement. The proposed approach can produce images with image restoration quality by including a standard image input and here, the neural network is evaluated by using images and features, which are obtained from separate areas to prove its capacity to generalize. The efficiency of the proposed approach is high when compared to other existing methods © 2021 IEEE.

关键词： image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Automated Defect Detection and Segregation in Dyed Fabrics Using image processing Techniques

Automated Defect Detection and Segregation in Dyed Fabrics U...

引用

Electronics and Sustainable Communication Systems (ICESC), 2020 International Conference on

作者： K. R. Prabha Aathavan. V Harshit. K Indhumathi. K Department of ECE Sri Ramakrishna Engineering College Coimbatore India

ISBN: (数字)9798350379945

ISBN: (纸本)9798350379952

This study addresses the critical need for efficient quality control in textile manufacturing, aiming to enhance the production of high-quality fabric to meet consumer demands and industry benchmarks. Despite the meticulous cleaning and dyeing processes involved in fabric production, there remains a challenge in identifying and segregating defective sections, which can potentially compromise the overall quality of the fabric. To overcome this challenge, this research study proposes an innovative solution: Automated Quality Control in Textile Manufacturing, utilizing advanced image processing techniques. The proposed system employs high-resolution cameras and image processing algorithms to analyze the fabric's surface in real-time. By leveraging predefined criteria and machine learning models, the system can accurately detect and isolate flawed sections within the fabric. These criteria encompass a range of defects, including stains, discolorations, or irregular patterns that deviate from established quality standards. This approach improves quality control in textile manufacturing, providing manufacturers with a reliable and efficient means of identifying and addressing defects in fabric production. By automating the quality control process, manufacturers can enhance productivity, reduce waste, and ultimately deliver higher-quality products to consumers.

关键词： Productivity image processing Process control Quality control Fabrics real-time systems Manufacturing Reliability Surface treatment Standards

来源：评论

学校读者我要写书评

暂无评论

deep-learning image Reconstruction for real-time Photoacoustic System

引用

IEEE TRANSACTIONS ON MEDICAL IMAGING 2020年第11期39卷 3379-3390页

作者： Kim, MinWoo Jeng, Geng-Shi Pelivanov, Ivan O'Donnell, Matthew Univ Washington Dept Bioengn UWAMIT Ctr Seattle WA 98105 USA Natl Chiao Tung Univ Dept Elect Engn Hsinchu 30010 Taiwan

Recent advances in photoacoustic (PA) imaging have enabled detailed images of microvascular structure and quantitative measurement of blood oxygenation or perfusion. Standard reconstruction methods for PA imaging are based on solving an inverse problem using appropriate signal and system models. For handheld scanners, however, the ill-posed conditions of limited detection view and bandwidth yield low image contrast and severe structure loss in most instances. In this paper, we propose a practical reconstruction method based on a deep convolutional neural network (CNN) to overcome those problems. It is designed for real-time clinical applications and trained by large-scale synthetic data mimicking typical microvessel networks. Experimental results using synthetic and real datasets confirm that the deep-learning approach provides superior reconstructions compared to conventional methods.

关键词： Photoacoustic imaging deep learning reconstruction convolutional neural network

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：