patternrecognition is an important way to improve the quality of life and promote scientific research. through the big data captured by the three-axis sensor, this paper focuses on the establishment of relevant model...
patternrecognition is an important way to improve the quality of life and promote scientific research. through the big data captured by the three-axis sensor, this paper focuses on the establishment of relevant models for the accurate classification of 19 motion modes , and analyses of the generalization ability and overfitting. this paper establishes DeepConvLSTM model by combining the original deep learning convolutional neural network (CNN) and long- term short-term memory recursion (LSTM). the model code about DeepConvLSTM was written by Python and relevant data was imported. In terms of methods, this paper evaluates the generalization ability of the model by using setaside method and K-fold cross-validation method respectively from the algorithm level. In terms of indicators, this paper selects accuracy, accuracy rate, recall rate and F1 value. Finally, the conclusion is drawn as follows: when epoch increases to 40, Train loss and Val loss basically fall together, indicating that the accuracy and stability of the model are *** paper use logistic regression model to solve the classification problem of human behavior. It is found that the classification effect of DeepConvLSTM model was better than that of LR model with higher accuracy and no overfitting phenomenon. in addition, Recall is used to analyze the sensitivity of the model. After several tests, it is found that the model shows good sensitivity and stability when the training times reaches 60. On this basis, considering the cost and efficiency of training comprehensively, the number of training is set as 60 in this paper, and the sensitivity and robustness of the model are both good. In conclusion, this paper establishes a scientific action patternrecognition model according to the conditions of the subject and the data given. the model is simple and easy to popularize. After verification and analysis, the model in this paper has strong accuracy, robustness and sensitivity, and has certain pra
this paper examines the technical aspects of utilizing Internet of Energy (IoE) metadata in smart city environments. through comparative and inductive analysis of existing literature and technical reports, we explore ...
详细信息
ISBN:
(数字)9798331532178
ISBN:
(纸本)9798331532185
this paper examines the technical aspects of utilizing Internet of Energy (IoE) metadata in smart city environments. through comparative and inductive analysis of existing literature and technical reports, we explore how advanced metadata management techniques can address inefficiencies in processing and analyzing large volumes of energy data in urban systems. Key findings include the potential for standardized metadata ontologies, edge computing for data preprocessing, and machine learning algorithms for patternrecognition in energy consumption. We propose strategies for developing unified IoE metadata management systems that integrate with broader smart city infrastructure. While metadata-driven approaches show promise for enhancing energy efficiency and sustainability in urban environments, challenges in data privacy and security remain. this research highlights IoE metadata as a critical tool for optimizing energy systems in smart cities through advanced data analytics and management techniques.
Electrical status epilepticus during sleep (ESES) is a specific EEG phenomenon induced by sleep with near-constant spike and slow wave emission. In the clinical diagnosis of its associated syndromes, the quantificatio...
Electrical status epilepticus during sleep (ESES) is a specific EEG phenomenon induced by sleep with near-constant spike and slow wave emission. In the clinical diagnosis of its associated syndromes, the quantification of abnormal electroencephalography (EEG) during sleep, i.e. spike-wave index (SWI), is often used as an important reference standard. It is based on the characteristics of spikes and slow-spikes in the central temporal region. EEG signals can identify the presence of seizures for patients and provide information about the severity and extent of abnormal EEG activity during sleep, which helps doctors to better understand the patient’s physiological function and to develop individualised treatment and prognosis plans. In this paper, we proposed a novel method that combines deep learning and morphological operations to identify and quantify epileptic electrical sustained activity during sleep. the proposed method provides the mean SWI error of 6.04%, the Recall of 87.37% and the Precision of 56.11%. Besides, 15.2%, 64.6% and 85.3% could be achieved for PCT (1%), PCT (5%) and PCT (10%), respectively. the experimental results show that the proposed method has great potential for the clinical diagnosis and prognosis of children with epilepsy. It will help to provide long-range EEG detection for patients with ESES syndrome, thus offering the possibility of early treatment for patients.
the use of convolutional neural networks (CNN) in the preservation of cultural heritage monuments, especially in conflict-affected regions such as Gaza, Ukraine, Iraq and others, represents a significant advancement i...
详细信息
ISBN:
(数字)9798350350265
ISBN:
(纸本)9798350350272
the use of convolutional neural networks (CNN) in the preservation of cultural heritage monuments, especially in conflict-affected regions such as Gaza, Ukraine, Iraq and others, represents a significant advancement in heritage conservation efforts. this paper presents an approach that uses a Multi-CNN model to classify images of cultural heritage monuments into various categories, encompassing period, monument type and location. By leveraging the capabilities of CNNs, this model demonstrates a high level of accuracy in categorizing heritage monuments based on multiple attributes. the study highlights the superior performance of the Multi-CNN model compared to other popular models such as DenseNet169, GoogleNet and MnasNet, highlighting its effectiveness in accurately classifying images of cultural heritage monuments in various dimensions. According to the evaluation results, the top-performing multi-CNN model achieves a classification accuracy of 94.52%, outperforming the single CNN models. the DenseNet196 model achieves 93.70% accuracy, the MnasNet model achieves 92.80% accuracy, and the GoogleNet model achieves 88.18% accuracy.
Table detection and structure recognition from archival document images remain challenging due to diverse table structures, complex document layouts, degraded image qualities and inconsistent table scales. In this pap...
详细信息
ISBN:
(纸本)9783030865498
Table detection and structure recognition from archival document images remain challenging due to diverse table structures, complex document layouts, degraded image qualities and inconsistent table scales. In this paper, we propose an instance segmentation based approach for archival table structure recognition which utilizes both foreground cell content and background ruling line information. To overcome the influence from inconsistent table scales, we design an adaptive image scaling method based on average cell size and density of ruling lines inside each document image. Different from previous multi-scale training and testing approaches which usually slow down the speed of the whole system, our adaptive scaling resizes each image to a single optimal size which can not only improve overall model performance but also reduce memory and computing overhead on average. Extensive experiments on cTDaR 2019 Archival dataset show that our method can outperform the baselines and achieve new state-of-the-art performance, which demonstrates the effectiveness and superiority of the proposed method.
Identifying the discharge type based on partial discharge signals in Gas-insulated Switchgear (GIS) equipment is very important for fault diagnosis and early warning of GIS equipment. this paper proposes a method to r...
详细信息
Normally, a patternrecognition has been studied in the field of digital signal processing with digital codes. Even though it has very accurate results, it needs a lot of power consumption, very long operating time, a...
详细信息
Normally, a patternrecognition has been studied in the field of digital signal processing with digital codes. Even though it has very accurate results, it needs a lot of power consumption, very long operating time, and huge hardware bundles. In this paper, an edge detected CMOS image sensor for an intelligent patternrecognition is proposed. Further, an efficient edge detection algorithm is also described. Withthe new techniques, a low power and high speed patternrecognition is available in the field of CMOS image sensor. the image sensor is composed of photodiodes, a two-step single-slope 8-bit analog-to-digital converter, static random memories, and etc. the testing sensor has been implemented with an 180nm technology. It has a very excellent measured results, compared to other published chips.
In problems of object and signal recognition, each of the errors of the first and second kind has its own cost, which takes on non-negative values. If they are equal, then the problem is relatively easy to solve. Sinc...
详细信息
this research paper aims to enhance the accuracy and efficiency of emotion recognitionthrough keystroke dynamics by utilizing a multi-class XGBoost model. the study addresses the limitations observed in previous mode...
详细信息
ISBN:
(数字)9798331517878
ISBN:
(纸本)9798331517885
this research paper aims to enhance the accuracy and efficiency of emotion recognitionthrough keystroke dynamics by utilizing a multi-class XGBoost model. the study addresses the limitations observed in previous models, such as Support Vector Machines and Logistic Regression, which struggled withthe complex, high-dimensional nature of keystroke data. the primary objective was to develop a robust model capable of capturing the subtle variations in typing behavior associated with different emotional states. To achieve this, a comprehensive approach was adopted, involving the collection of detailed keystroke datasets, advanced feature engineering, and systematic hyperparameter tuning. the XGBoost model was chosen for its superior ability to manage complex interactions within large datasets. through rigorous evaluation, the XGBoost model outperformed traditional models in accuracy, precision, and recall, demonstrating significant improvements in emotion classification tasks. the findings of this study reveal that XGBoost, when combined with meticulous feature engineering and hyperparameter optimization, offers a substantial advancement in the field of emotion recognition via keystroke dynamics. the research not only confirms the effectiveness of XGBoost in this domain but also provides new insights into optimizing machine learning models for emotion detection. these results have broad implications for the development of more emotionally aware computing systems, potentially improving user interaction across various applications.
this paper proposes a novel framework for accurately recognizing physical exergames interaction in video sequences. the developed approach uses image processing and machine learning tools to correctly extract and clas...
详细信息
ISBN:
(数字)9798331533038
ISBN:
(纸本)9798331533045
this paper proposes a novel framework for accurately recognizing physical exergames interaction in video sequences. the developed approach uses image processing and machine learning tools to correctly extract and classify the features. the framework comprises five main steps: Silhouette extraction, preprocessing, feature extraction, feature optimization and classification. Image contrast is improved by power law transformation, silhouette is extracted by the Multiple Object Tracking (MOT) algorithm and graph-based segmentation. We extract ORB and geometric skeleton based keypoints to capture discriminative features, and apply Fast Independent Component Analysis (FICA) to optimize feature representation. thirdly, the Convolutional Neural Network algorithm is used to classify the optimized features, with a macro average accuracy of 0.85 and a weighted average accuracy of 0.87%. the proposed framework effectively recognizes physical exergames interaction and can be applied in many domains, such as human-human interaction and surveillance.
暂无评论