检索结果-内蒙古大学图书馆

Novel methodology for apple leaf disease classification with PCNN-IELM

Neural Computing and Applications 2025年第6期37卷 4895-4913页

作者： Navpreet Roul, Rajendra Kumar Department of Computer Science and Engineering Thapar Institute of Engineering and Technology Punjab Patiala147004 India

Agriculture is crucial to the global economy, particularly in ensuring food security. Recent trends indicate that various plant diseases are causing substantial financial losses in the agricultural sector worldwide. Traditional manual inspection methods for detecting fruit and plant diseases are labor-intensive and inefficient. Adopting automated disease detection technologies could significantly enhance early diagnosis and reduce the economic impact of these diseases on agriculture. This study introduces an advanced model for classifying apple diseases by integrating a pre-trained convolutional neural network (PCNN), such as VGG16, VGG19, or ResNet50, with an incremental extreme learning machine (I-ELM) for efficient feature extraction and classification. A key innovation of this model is replacing the PCNN’s fully connected layer with the I-ELM, which eliminates the lengthy back-propagation process and significantly reduces training time. Integrating I-ELM with PCNN harnesses the rapid learning capabilities and robust generalization of I-ELM with the superior feature extraction abilities of CNNs. I-ELM simplifies the network architecture by avoiding the complex neural networks commonly used in other methods. The model’s effectiveness is rigorously evaluated on the well-known Plant Village dataset, demonstrating its ability to identify various apple diseases through performance metrics such as precision, sensitivity, specificity, accuracy, and the F1-score. Comparing existing deep learning models using these metrics highlights its superior performance. This innovation is up-and-coming for intelligent agricultural systems, offering an effective solution for classifying apple diseases and enabling timely and innovative farming practices. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Fruits

来源：评论

学校读者我要写书评

暂无评论

Cultivating road safety: A comprehensive examination of intelligent ensemble-based road crack detection

引用

Multimedia Tools and Applications 2025年第12期84卷 10747-10780页

作者： Navpreet Roul, Rajendra Kumar Rani, Rinkle Department of Computer Science and Engineering Thapar Institute of Engineering & Technology Punjab Patiala147004 India

Detecting and promptly identifying cracks on road surfaces is of paramount importance for preserving infrastructure integrity and ensuring the safety of road users, including both drivers and pedestrians. Presently, the predominant approach for crack detection relies on labor-intensive and costly manual inspections, exacerbated by the enduring challenge of concrete road cracks susceptible to environmental factors like temperature fluctuations, heavy traffic loads, and prolonged exposure to harsh weather conditions. This paper presents a comprehensive research initiative aimed at advancing road crack detection techniques. The proposed method leverages a classical Convolutional Neural Network (CNN) in conjunction with the Extreme Learning Machine (ELM) for efficient feature extraction and classification. A pivotal breakthrough is achieved by replacing the fully connected layer of the CNN with ELM, thus circumventing the time-consuming backpropagation process and significantly expediting the training process. This integration harnesses ELM’s swift learning capabilities and strong generalization performance, complemented by the CNN’s exceptional feature extraction abilities. The model’s effectiveness is rigorously assessed using the wide recognized CCIC, SDNET2018, and Synthesized datasets, and its proficiency in distinguishing between Crack and No-Crack regions is demonstrated through various performance metrics. Furthermore, the model’s performance is benchmarked against existing deep learning models using these performance metrics, showcasing its superior performance. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Roads and streets

来源：评论

学校读者我要写书评

暂无评论

A Novel CAPTCHA Recognition System Based on Refined Visual Attention

引用

computers, Materials & Continua 2025年第4期83卷 115-136页

作者： Zaid Derea Beiji Zou Xiaoyan Kui Monir Abdullah Alaa Thobhani Amr Abdussalam School of Computer Science and Engineering Central South UniversityChangsha410083China College of Computer Science and Information Technology Wasit UniversityWasit52001Iraq Department of Computer Science and Artificial Intelligence College of Computing and Information TechnologyUniversity of BishaBisha67714Saudi Arabia Electronic Engineering and Information Science Department University of Science and Technology of ChinaHefei230026China

Improving website security to prevent malicious online activities is crucial,and CAPTCHA(Completely Automated Public Turing test to tell computers and Humans Apart)has emerged as a key strategy for distinguishing human users from automated ***-based CAPTCHAs,designed to be easily decipherable by humans yet challenging for machines,are a common form of this ***,advancements in deep learning have facilitated the creation of models adept at recognizing these text-based CAPTCHAs with surprising *** our comprehensive investigation into CAPTCHA recognition,we have tailored the renowned UpDown image captioning model specifically for this *** approach innovatively combines an encoder to extract both global and local features,significantly boosting the model’s capability to identify complex details within CAPTCHA *** the decoding phase,we have adopted a refined attention mechanism,integrating enhanced visual attention with dual layers of Long Short-Term Memory(LSTM)networks to elevate CAPTCHA recognition *** rigorous testing across four varied datasets,including those from Weibo,BoC,Gregwar,and Captcha 0.3,demonstrates the versatility and effectiveness of our *** results not only highlight the efficiency of our approach but also offer profound insights into its applicability across different CAPTCHA types,contributing to a deeper understanding of CAPTCHA recognition technology.

关键词： Text-based CAPTCHA recognition refined visual attention web security computer vision

来源：评论

学校读者我要写书评

暂无评论

Leveraging feature fusion ensemble of VGG16 and ResNet-50 for automated potato leaf abnormality detection in precision agriculture

引用

Soft Computing 2025年第4期29卷 2263-2277页

作者： Trivedi, Amit Kumar Mahajan, Tripti Maheshwari, Tanmay Mehta, Rajesh Tiwari, Shailendra Department of Computer Science and Engineering Thapar Institute of Engineering and Technology Punjab Patiala147001 India

In the era of advancement in technology and modern agriculture, early disease detection of potato leaves will improve crop yield. Various researchers have focussed on disease due to different types of microbial infection in potato leaves using computer vision and machine learning approaches. In this paper, a data science approach for multiclass classification of potato normal and abnormal leaves due to fungal infection like early blight and late blight is performed using the ensembling of deep learning (DL) CNN models. Firstly, the performance of classification on potato disease is verified separately on VGG16 and ResNet-50 CNN models after pre-processing of the leaf dataset. The pre-processing includes noise removal and normalization. Further improvement in classification accuracy is achieved by the ensembling of VGG16 and ResNet-50 CNN models. The ensembling of CNN models is performed on the feature level by fusing features extracted using VGG16 and ResNet-50. From the experimental results, performed on publicly available datasets consisting of 2152 number of normal and abnormal images it is observed that the average classification accuracy of 98.22%, 96.16% and 95.68% is achieved using the proposed ensemble, VGG16 and ResNet-50 models respectively. The efficacy of the proposed approach (ensemble technique at feature level fusion) is verified in comparison with recently reported DL model-based approaches. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

RoleNet: A multiple features fusion network for role classification in cantonese opera

引用

Multimedia Tools and Applications 2025年 1-14页

作者： Li, Yue Peng, Zhengwei Xu, Di Chen, Yuanguang Chen, Guoan School of Computer Science and Engineering South China University of Technology Guangdong Guangzhou510006 China School of Computer Science and Engineering Sun Yat-sen University Guangdong Guangzhou510006 China

Cantonese opera, a key facet of Chinese traditional opera, boasts profound cultural and artistic value and has been designated as intangible cultural heritage. The use of certain roles is a basic concept in Cantonese opera, where each role has a specific style of singing, movement, and costume that performers are trained to perform throughout their careers. Therefore, identifying the role category of characters in a play can provide theoretical and systematic foundations for further researches and artistic explorations. By dissecting musical traits and performance styles of each role, comprehensive studies on its regional and artistic nuances are enabled. To achieve role classification in Cantonese opera, we propose RoleNet, an integration network that consists of SincNets, transformers, and a feature fusion block. For a given musical fragment (e.g., an audio signal), SincNets extract 1D features at multiple scales and transformers extract features from time and frequency axes from the 2D Mel Spectrogram. Subsequently, the extracted features are concatenated by the fusioner using multiple features selection strategy to perform role classification tasks. The experimental results on a real-world dataset demonstrated the superior performance of RoleNet compared to single-objective methods. An overall classification accuracy of 98.67% was achieved. The codes are available at https://***/AaronPeng920/RoleNet. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Feature Selection

来源：评论

学校读者我要写书评

暂无评论

Enhancing English accent identification in automatic speech recognition using spectral features and hybrid CNN-BiLSTM model

引用

Multimedia Tools and Applications 2025年 1-28页

作者： Ahmed, Ghayas Lawaye, Aadil Ahmad Jain, Vishal Chatterjee, Jyotir Moy Mahajan, Shubham Department of Computer Sciences Baba Ghulam Shah Badshah University J&K Rajouri India Department of Computer Science & Engineering Sharda University Noida India Dehradun India Department of Computer Science & Engineering Amity University Haryana India

Automatic Speech Recognition (ASR) has been the regnant research area in the domain of Natural Language Processing for the last few decades. Past years’ advancement provides progress in this area of research. The accent of spoken is the prominent factor affecting the performance of ASR. Accent is the pronunciation of a word in a distinct way which may vary from specimen to specimen depending upon their social class, formality, age, vocal property, gender, geography, and influence of native or other language. Recognizing an accent requires various complex characteristics and features of voice such as voice quality, prosody, and phoneme pronunciation which are difficult to extract and analyze. To solve these difficulties, the researchers focus their attention on spectral representation such as Mel Spectrogram and Mel-Frequency Cepstral Coefficient (MFCC). In this study, we will analyze which features achieve maximum accuracy for the English accent classification task. Here in this work, we perform our experiments on Log Mel filter bank and MFCC features with four different window function. We extract features from raw audio with various windowing function and train a hybrid model of Convolutional Neural Network and Bidirectional Long Short-term Memory network (CNN-BiLSTM) on a custom dataset having nine different accents of English language namely American, Australian, British, Indian (Oriya, Bangla, Telegu, Malayalam), and Welsh. The accuracy of each feature is evaluated and compared. The log Mel filter bank outputs the highest accuracy of 99.75% and 99.91% of training and validation with Han window function while MFCC with Bartlett window function achieves 99.98% of training and 99.33% of validation accuracy. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Spectrographs

来源：评论

学校读者我要写书评

暂无评论

Composite spectral spatial pixel CNN for land-use hyperspectral image classification with hybrid activation function

引用

Multimedia Tools and Applications 2025年第12期84卷 10527-10550页

作者： Banerjee, Anasua Swain, Satyajit Rout, Minakhi Bandyopadhyay, Mainak School of Computer Engineering Kalinga Institute of Industrial Technology-Deemed-to-be-University Bhubaneswar India Department of Computer Science and Engineering National Institute of Technology Silchar India

Deep learning methods have played a prominent role in the development of computer visualization in recent years. Hyperspectral imaging (HSI) is a popular analytical technique based on spectroscopy and visible imaging which examines the pattern of light in a target and recognizes objects based on varying spectral properties. However, in remote sensing, detecting surface material via HSI analysis is a critical and difficult task. The performance of spectral-spatial data exploitation is well established to outperform typical spectral pixel-wise techniques. Because of its great feature extraction ability, convolutional neural networks (CNN) have emerged as a potent deep learning approach. CNN translates the input features of an image into an equivalent CNN feature map, in addition to naturally combining spectral and spatial information. However, spectral-spatial properties when combined with pixel-wise extraction can learn more minute details of objects present on the earth’s terrain. In this paper, a noble Composite Spectral Spatial Pixel CNN model for the classification of hyperspectral data is presented which is an amalgamation of 3D-2D-1D CNN. While the 3D and 2D CNN exploit the spectral-spatial features effectively, 1D CNN works on pixel-wise feature extraction. Further, to optimize the classification performance of the proposed model, a new hybrid activation function Flatten-T Swish is also used which is the combination of ReLU and Swish function. The proposed model is compared with other state-of-the-art models based on three popular HSI datasets, and it is found that the proposed model performs better among others in terms of classification and computation time, giving 98.87% accuracy for Indian Pines, 99.92% accuracy for Pavia University, and 99.99% accuracy for Salinas Valley dataset. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Extraction

来源：评论

学校读者我要写书评

暂无评论

Dual-Task Contrastive Meta-Learning for Few-Shot Cross-Domain Emotion Recognition

引用

computers, Materials & Continua 2025年第2期82卷 2331-2352页

作者： Yujiao Tang Yadong Wu Yuanmei He Jilin Liu Weihan Zhang School of Computer Science and Engineering Sichuan University of Science and EngineeringYibin644002China School of Mechanical and Power Engineering Chongqing University of Science and TechnologyChongqing401331China

Emotion recognition plays a crucial role in various fields and is a key task in natural language processing (NLP). The objective is to identify and interpret emotional expressions in text. However, traditional emotion recognition approaches often struggle in few-shot cross-domain scenarios due to their limited capacity to generalize semantic features across different domains. Additionally, these methods face challenges in accurately capturing complex emotional states, particularly those that are subtle or implicit. To overcome these limitations, we introduce a novel approach called Dual-Task Contrastive Meta-Learning (DTCML). This method combines meta-learning and contrastive learning to improve emotion recognition. Meta-learning enhances the model’s ability to generalize to new emotional tasks, while instance contrastive learning further refines the model by distinguishing unique features within each category, enabling it to better differentiate complex emotional expressions. Prototype contrastive learning, in turn, helps the model address the semantic complexity of emotions across different domains, enabling the model to learn fine-grained emotions expression. By leveraging dual tasks, DTCML learns from two domains simultaneously, the model is encouraged to learn more diverse and generalizable emotions features, thereby improving its cross-domain adaptability and robustness, and enhancing its generalization ability. We evaluated the performance of DTCML across four cross-domain settings, and the results show that our method outperforms the best baseline by 5.88%, 12.04%, 8.49%, and 8.40% in terms of accuracy.

关键词： Contrastive learning emotion recognition cross-domain learning dual-task meta-learning

来源：评论

学校读者我要写书评

暂无评论

AmplitudeArrow: On-the-Go AR Menu Selection Using Consecutive Simple Head Gestures and Amplitude Visualization

引用

IEEE Transactions on Visualization and computer Graphics 2025年第05期31卷 3408-3417页

作者： Tian, Yang Zhang, Youpeng Yan, Yukang Zhao, Shengdong Ma, Xiaojuan Shi, Yuanchun Guangxi University Department of Computer Science Guangxi Key Laboratory of Multimedia Communications and Network Technology China University of Rochester Department of Computer Science United States City University of Hong Kong School of Creative Media and the Department of Computer Science Hong Kong Hong Kong University of Science and Technology Department of Computer Science and Engineering Hong Kong Tsinghua University Department of Computer Science and Technology China

Heads-up computing aims to provide synergistic digital assistance that minimally interferes with users' on-the-go daily activities. Currently, the input modalities of heads-up computing are mainly voice and finger gestures. In this work, we propose and evaluate the AmplitudeArrow (AA) technique designed for on-the-go AR menu selection to demonstrate that consecutive simple head gestures can also be an effective input modality for heads-up computing. Specifically, AA arranges menu icons into one/two row(s). To select a target icon, the user first makes their head yaw to pre-select the target icon or the column containing it and then makes their head pitch to make the arrow in the target icon expand until the arrow covers the target icon completely, i.e., the pitch amplitude surpasses the selection confirmation threshold. User studies indicated that AA demonstrated robust resistance to walking-caused head perturbation and external factors such as other people/obstacles, delivering high accuracy (error rate © 1995-2012 IEEE.

关键词： Augmented reality

来源：评论

学校读者我要写书评

暂无评论

Text-driven clothed human image synthesis with 3D human model estimation for assistance in shopping

引用

Multimedia Tools and Applications 2025年第1期84卷 167-200页

作者： Karkuzhali, S. Aasim, A. Syed StalinRaj, A. Department of Computer Science and Engineering Mepco Schlenk Engineering College Tamil Nadu Sivakasi India

Online shopping has become an integral part of modern consumer culture. Yet, it is plagued by challenges in visualizing clothing items based on textual descriptions and estimating their fit on individual body types. In this work, we present an innovative solution to address these challenges through text-driven clothed human image synthesis with 3D human model estimation, leveraging the power of Vector Quantized Variational AutoEncoder (VQ-VAE). Creating diverse and high-quality human images is a crucial yet difficult undertaking in vision and graphics. With the wide variety of clothing designs and textures, existing generative models are often not sufficient for the end user. In this proposed work, we introduce a solution that is provided by various datasets passed through several models so the optimized solution can be provided along with high-quality images with a range of postures. We use two distinct procedures to create full-body 2D human photographs starting from a predetermined human posture. 1) The provided human pose is first converted to a human parsing map with some sentences that describe the shapes of clothing. 2) The model developed is then given further information about the textures of clothing as an input to produce the final human image. The model is split into two different sections the first one being a codebook at a coarse level that deals with overall results and a fine-level codebook that deals with minute detailing. As mentioned previously at fine level concentrates on the minutiae of textures, whereas the codebook at the coarse level covers the depictions of textures in structures. The decoder trained together with hierarchical codebooks converts the anticipated indices at various levels to human images. The created image can be dependent on the fine-grained text input thanks to the utilization of a blend of experts. The quality of clothing textures is refined by the forecast for finer-level indexes. Implementing these strategies can result

关键词： Vector quantization

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：