检索结果-内蒙古大学图书馆

A Comprehensive Review On Two-Stage Object Detection Algorithms

A Comprehensive Review On Two-Stage Object Detection Algorit...

2023 IEEE International Conference on Quantum Technologies, Communications, computing, Hardware and Embedded Systems Security, iQ-CCHESS 2023

作者： Varghese, Rejin Sambath, M. Hindustan Institute of Technology and Science Department of Computer Applications Chennai India Hindustan Institute of Technology and Science School of Computing Science Chennai India

ISBN: (纸本)9798350314946

Two-stage object detection algorithms have gained significant attention in computer vision due to their robustness and accuracy in detecting objects in images. These algorithms consist of two main stages: region proposal generation and object classification/localization. The region proposal stage efficiently generates a diverse set of candidate object regions using various techniques, narrowing down the search space. In the second stage, deep convolutional neural networks extract discriminative features from the proposed regions, enabling precise object classification and localization. These algorithms have demonstrated impressive performance in complex scenes with occlusions, scale variations, and multiple objects. However, they have limitations such as slower inference speeds, increased computational complexity, and the need for additional training steps. To address these limitations, more efficient algorithms like single-stage detectors have emerged. Two-stage object detection algorithms offer a powerful framework for accurate and robust object detection, and ongoing research aims to refine and improve these algorithms, expanding the capabilities of computer vision applications. © 2023 IEEE.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Handwritten optical character recognition using TransRNN trained with self improved flower pollination algorithm (SI-FPA)

引用

Multimedia Tools and Applications 2025年第18期84卷 19947-19969页

作者： Devi, S. Nandhini Fatima, N. Sabiyath Department of Computer Science and Engineering B.S. Abdur Rahman Crescent Institute of Science & Technology Chennai India

The difficulty of successfully scanning handwritten text arises from variances in style, size, and orientation, which affect handwriting optical character recognition (OCR). This study provides a novel strategy that incorporates innovative techniques for effectively addressing these difficulties. Initially, preprocessing steps involve Gaussian blur to reduce noise and CLAHE to enhance contrast, improving the visibility of text regions. Canny Edge Detection is then utilized for segmentation, followed by text localization using the Efficient and Accurate Scene Text (EAST) algorithm. Next, features are extracted via Inception V3 to capture both low-level and high-level information. From the collected characteristics, Long Short-Term Memory (LSTM) networks are used to extract chronological data. Then, with the help of the Self-Improved Flower Pollination Algorithm (SI-FPA), features are selected using Optimized Principal Component Analysis (PCA). Finally, a combination of Transformer and Recurrent neural networks (RNNs) is used to perform text recognition, maximizing the benefits of each network’s advantages in processing sequential and contextual data. Experiments on MNST dataset with validation using different optimization algorithms show the effectiveness and resilience of the suggested method in different optimization scenarios. The combination of these methods results in a sophisticated OCR system that can recognize handwritten characters with accuracy, even under difficult conditions. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Automatic Brain Tumor Detection Classification Applying Multiple a Deep Learning Method

Automatic Brain Tumor Detection Classification Applying Mult...

引用

2024 International Conference on System, Computation, Automation and Networking, ICSCAN 2024

作者： Vijiyakumar, K. Muthunagai, R. Bhanusree, A. Nikhilsowrabh, K. School of Computing Srm Institute of Science and Technology Department of Computing Technology Chenngalpet district Kattankulathur India Manakula Vinayagar Institute of Technology Department of Electrical & Electronics Engineering Puducherry India Srm Institute of Science and Technology Department of Computing Technologies Chennai India

ISBN: (纸本)9798331510022

Automated detection of defects in medical imaging has emerged as a crucial domain within various diagnostic applications, especially with MRI regarding tumor detection. 'While accuracy in the detection of tumors is important in providing vital information regarding tissues which have deviated from their normal state, it's even more important for effective treatment planning'. Currently, the diagnosis of tumors from brain MRI scans relies on human assessment, which is infeasible for large image data and error-prone based on judgments of humans. Thus, there is a pressing need for robust and efficient automated classification systems that can overcome these bottlenecks. The core idea of our project is the development of the most advanced tumor detection methods, which saves precious time for radiologists but at a very high accuracy and, thus, reduces mortality levels in patients. In our approach, we have used several CNNs - VGG19, ResNet50, and MobileNetV2, which we have applied through transfer learning for detecting tumors from brain MRIs. We also incorporate into our model an ensemble of these varied CNN architectures that leverages their strengths since the nature and diversity of tumors are intricate. In fact, the ensemble method is really intended to boost the wide-ranging accuracy and strength of the procedure in the detection of tumors. The proposed system utilizes deep learning techniques in combination with image classification to accurately and efficiently identify cancerous cells in brain MRI scans. After that, the web application, developed using Flask, will provide a user-friendly interface for uploading MRI images and then instantaneously return predictions to a doctor on the potential presence of a tumor. © 2024 IEEE.

关键词： Convolutional Neural Networks (CNNs) Deep Learning EnsembleMethod Flask Web Application MRI Imaging TransferLearning Tumor Detection

来源：评论

学校读者我要写书评

暂无评论

IoT-driven sport activity recognition system empowered by deep learning and wearable technology 4th

IoT-driven sport activity recognition system empowered by de...

引用

4th International Conference on Computational Methods in science and technology, ICCMST 2024

作者： Giridhar, Koppisetti Hyndhavi, Dharmavaram Hayath, Shaik Nayab Mohammed Devichandu, Kamsala Khan, Deshmukh Faris Department of Computer Science and Technology Madanapalle Institute of Technology and Science Andhra Pradesh India

ISBN: (纸本)9781032911571

This article affords an innovative type of system for wearable sporting sports that utilises a deep studying algorithm to accurately locate various sports. Two inertial sensor modules are covered within the machine, which athletes wear on their wrists and ankles to record movement of sporting activities. By employing a deep convolutional neural community (CNN). Each inertial sensing module consists of a triaxial accelerometer, microcontroller, triaxial gyroscope, RF wireless transmission module. Furthermore, a website has been advanced using HTML, CSS, and JavaScript, allowing athletes to examine their overall performance. The proposed deep getting to know-based class algorithm encompasses various steps such as CNN-primarily based type, photograph resizing, spectrogram era, normalisation, movement segmentation, and signal processing, enabling the identification of ten forms of athletic endeavours. The cautioned machine and algorithm are successful in successfully figuring out a number of athletic sports, as established by means of the effects of the experiments. © 2025 the Author(s).

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Deoxyribonucleic Acid Based Oryza Sativa Phenotype Prediction Using Molecular Neural Network

Deoxyribonucleic Acid Based Oryza Sativa Phenotype Predictio...

引用

2024 International Conference on Advances in Data Engineering and Intelligent computing Systems, ADICS 2024

作者： Soren, Nikita Rajendran, P. Selvi School Of Computing Science Hindustan Institute Of Technology And Science Chennai India

ISBN: (纸本)9798350364828

Precise phenotypic trait prediction in Oryza sativa (rice) is essential for breeding initiatives, agricultural innovations, and maintaining food security. In this work, we present a unique method for phenotype prediction based on the examination of Oryza sativa chromosomes 1, 2, 4, 7, 11, and 12 deoxyribonucleic acid (DNA) sequences. In order to capture the complex genetic properties linked to phenotypic variances, we created a Molecular Neural Network (MNN) architecture specifically designed for processing molecular data. This architecture integrates both recurrent and convolutional layers. We preprocess the DNA sequences using one-hot encoding and data partitioning for testing, validation, and training as part of our technique. A dataset of DNA sequences and associated phenotypic labels is used to train the MNN. The effectiveness of the model is assessed using common measures like accuracy. The model's performance on a test set is being measured in the proposed study by computing Pearson correlation coefficients between expected and actual trait values. It also calculates the (PCC) for each feature in the model, and in 152 seconds, we obtained an accuracy of 98%. © 2024 IEEE.

关键词： DNA

来源：评论

学校读者我要写书评

暂无评论

Effective Scene Graph Generation by Statistical Relation Distillation

Effective Scene Graph Generation by Statistical Relation Dis...

引用

2025 IEEE/CVF Winter Conference on Applications of computer Vision, WACV 2025

作者： Nguyen, Thanh-Son Yang, Hong Fernando, Basura Institute of High Performance Computing Agency for Science Technology and Research Singapore Agency for Science Technology and Research Singapore

ISBN: (纸本)9798331510831

Annotating scene graphs for images is a time-consuming task, resulting in many instances of missing relations within existing datasets. In this paper, we introduce the Statistical Relation Distillation (SRD) method to enhance scene graph datasets. SRD leverages human-annotated relations alongside object-to-object and predicate-to-predicate similarities to reinforce the existence likelihood of scene graph relations. Moreover, SRD can augment relational frequency using relations of non-selected object and predicate categories that are usually omitted by scene graph generation (SGG) task. The output from SRD derives the prior probability which is combined with model-predicted probabilities to annotate missing relations in training images and subsequently re-train SGG models on the augmented dataset. We evaluate our proposed method on Visual Genome and GQA-200 datasets. Experimental results show that training on the augmented dataset enhances the performance of prominent scene-graph generation models. The implementation code is at https://***/LUNAProject22/SRD. © 2025 IEEE.

关键词： Visualization

来源：评论

学校读者我要写书评

暂无评论

The ringed residual u-net with non-natural regions feature for image splicing forgery detection and localization

引用

Journal of Intelligent and Fuzzy Systems 2024年第4期46卷 7447-7459页

作者： Wang, Qi Lu, Tongwei School of Computer Science and Engineering Wuhan Institute of Technology Wuhan China Hubei Key Laboratory of Intelligent Robot Wuhan Institute of Technology Wuhan China

Recently, with the emergence of many image editing tools (photoshop, Topaz studio, etc.), the authenticity of images has been severely challenged. However, the performance of some existing traditional feature extraction methods and detection methods based on convolutional neural network (CNN) is poor, and the information provided by the features extracted from the network is limited and single. In this paper, an end-To-end ringed residual U-Net is proposed to detect image splicing forgery by blending features of non-natural regions. Some regions with significant differences from the image background are defined as non-natural regions(such as the irregular border at the splicing of images). In this paper, a feature enhancement module for non-natural regions is constructed, which the image through the pooling of four different scales, and these features are then combined with the original image and input to the backbone network for processing, aiming to highlight regions of the image that differ significantly from the background. Therefore, after adding the feature enhancement module for non-natural regions to the end-To-end ring residual U-Net, more attention will be paid to the tampering regions in the feature extraction stage, image manipulation detection and localization will also become more accurate. Compared with some mainstream methods, this method achieves better performance on the three standard datasets(CASIA2.0, NIST2016, COLUMBIA). In addition, it has excellent robustness under JPEG compression attack and noise corruption attack. © 2024-IOS Press. All rights reserved.

关键词： Topaz

来源：评论

学校读者我要写书评

暂无评论

Feature Selection with Deep Belief Network for Epileptic Seizure Detection on EEG Signals

引用

computers, Materials & Continua 2023年第5期75卷 4101-4118页

作者： Srikanth Cherukuvada R.Kayalvizhi Department of Networking and Communications School of ComputingSRM Institute of Science and TechnologyKattankulathurChennai603203India

The term Epilepsy refers to a most commonly occurring brain disorder after a *** identification of incoming seizures significantly impacts the lives of people with *** detection of epileptic seizures(ES)has dramatically improved the life quality of the *** Electroencephalogram(EEG)related seizure detection mechanisms encountered several difficulties in *** EEGs are the non-stationary signal,and seizure patternswould changewith patients and recording ***,EEG data were disposed to wide noise varieties that adversely moved the recognition accuracy of *** intelligence(AI)methods in the domain of ES analysis use traditional deep learning(DL),and machine learning(ML)*** article introduces an Oppositional Aquila Optimizer-based Feature Selection with Deep Belief Network for Epileptic Seizure Detection(OAOFS-DBNECD)technique using EEG *** primary aim of the presented OAOFS-DBNECD system is to categorize and classify the presence of *** suggested OAOFS-DBNECD technique transforms the EEG signals *** format at the initial ***,the OAOFS technique selects an optimal subset of features using the preprocessed *** seizure classification,the presented OAOFS-DBNECD technique applies Artificial Ecosystem Optimizer(AEO)with a deep belief network(DBN)*** extensive range of simulations was performed on the benchmark dataset to ensure the enhanced performance of the presented OAOFS-DBNECD *** comparison study shows the significant outcomes of the OAOFS-DBNECD approach over other *** addition,the result of the suggested approach has been evaluated using the CHB-MIT database,and the findings demonstrate accuracy of 97.81%.These findings confirmed the best seizure categorization accuracy on the EEG data considered.

关键词： Seizure detection EEG signals machine learning deep learning feature selection

来源：评论

学校读者我要写书评

暂无评论

Hybrid Deep Learning Model to Predict Students’ Sentiments in Higher Educational Institutions

International Journal of Interactive Mobile Technologies

引用

International Journal of Interactive Mobile Technologies 2025年第1期19卷 46-61页

作者： Ananthi, Claral Mary T. Department of Computer Applications Faculty of Science and Humanities SRM Institute of Science and Technology Kattankulathur Tamil Nadu Chengalpattu India

Sentiment analysis has been widely used in various fields of social media, education, and business. Specifically, in the education domain, the usage of sentiment analysis is difficult due to the huge amount of information, the nature of language, and processing the diverse perceptions of students. Deep learning emerges as an advanced concept in the realm of machine learning that learns features automatically from raw text data, making them well-suited for sentiment analysis tasks. In recent years, deep learning has been used in analyzing the sentiments. Deep learning architectures have surpassed other machine learning paradigms for performing sentiment analysis. The ability to analyze automatically the students’ sentiments enables HEI to process huge amounts of unstructured data quickly, efficiently, and cost-effectively. The paper aims to predict the sentiments of students’ reviews posted in VLE regarding online learning that enables the educators to optimize their teaching methods for the best results. This study paper explores the usage of CNN, LSTM, and hybrid CNN-LSTM for the prediction of sentiments. The proposed hybrid CNN-LSTM architecture achieves superior performance compared to other baseline algorithms with respect to accuracy, precision, recall, and F1 score. According to outcomes, the recommended technique achieves remarkable accuracy of 97%. The findings facilitate the progress of a more efficient deep learning sentiment prediction system that gives valuable insights from a huge volume of students’ textual data. © 2025 by the authors of this article.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

A Computational Deep Fuzzy Network-Based Neuroimaging Analysis for Brain Hemorrhage Classification

引用

IEEE Journal of Biomedical and Health Informatics 2023年 PP卷 1-8页

作者： Malik, Payal Vidyarthi, Ankit Department of Computer Science Engineering and Information Technology Jaypee Institute of Information Technology Noida India

The boundaries and regions between individual classes in biomedical image classification are hazy and overlapping. These overlapping features make predicting the correct classification result for biomedical imaging data a difficult diagnostic task. Thus, in precise classification, it is frequently necessary to obtain all necessary information before making a decision. This paper presents a novel deep-layered design architecture based on Neuro-Fuzzy-Rough intuition to predict hemorrhages using fractured bone images and head CT scans. To deal with data uncertainty, the proposed architecture design employs a parallel pipeline with rough-fuzzy layers. In this case, the rough-fuzzy function functions as a membership function, incorporating the ability to process rough-fuzzy uncertainty information. It not only improves the deep model's overall learning process, but it also reduces feature dimensions. The proposed architecture design improves the model's learning and self-adaptation capabilities. In experiments, the proposed model performed well, with training and testing accuracies of 96.77% and 94.52%, respectively, in detecting hemorrhages using fractured head images. The comparative analysis shows that the model outperforms existing models by an average of 2.6$\pm$0.90% on various performance metrics. IEEE

关键词： Classification (of information)

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：