Parallel convolutional neural network and empirical mode decomposition for high accuracy in motor imagery EEG signal classification

PLoS ONE, Год журнала: 2025, Номер 20(1), С. e0311942 - e0311942

Опубликована: Янв. 16, 2025

In recent years, the utilization of motor imagery (MI) signals derived from electroencephalography (EEG) has shown promising applications in controlling various devices such as wheelchairs, assistive technologies, and driverless vehicles. However, decoding EEG poses significant challenges due to their complexity, dynamic nature, low signal-to-noise ratio (SNR). Traditional pattern recognition algorithms typically involve two key steps: feature extraction classification, both crucial for accurate operation. this work, we propose a novel method that addresses these by employing empirical mode decomposition (EMD) parallel convolutional neural network (PCNN) classification. This approach aims mitigate non-stationary issues, improve performance speed, enhance classification accuracy. We validate effectiveness our proposed using datasets BCI competition IV, specifically 2a 2b, which contain signals. Our focuses on identifying two- four-class signal classifications. Additionally, introduce transfer learning technique fine-tune model individual subjects, leveraging important features extracted group dataset. results demonstrate EMD-PCNN outperforms existing approaches terms conduct qualitative quantitative analyses evaluate method. Qualitatively, employ confusion matrices metrics specificity, sensitivity, precision, accuracy, recall, f1-score. Quantitatively, compare accuracies with those approaches. findings highlight superiority accurately classifying The enhanced robustness underscore its potential broader applicability real-world scenarios.

Язык: Английский

---

Human Posture Detection Using Image Augmentation and Hyperparameter-Optimized Transfer Learning Algorithms

Applied Sciences, Год журнала: 2022, Номер 12(19), С. 10156 - 10156

Опубликована: Окт. 10, 2022

With the advancement in pose estimation techniques, human posture detection recently received considerable attention many applications, including ergonomics and healthcare. When using neural network models, overfitting poor performance are prevalent issues. Recently, convolutional networks (CNNs) were successfully used for recognition from images due to their superior multiscale high-level visual representations over hand-engineering low-level characteristics. However, calculating millions of parameters a deep CNN requires significant number annotated examples, which prohibits CNNs such as AlexNet VGG16 being on issues with minimal training data. We propose new three-phase model decision support that integrates transfer learning, image data augmentation, hyperparameter optimization (HPO) address this problem. The is part framework hyperparameters AlexNet, VGG16, CNN, multilayer perceptron (MLP) models accomplishing optimal classification results. learning algorithms HPO detection, while Multilayer Perceptron standard classifiers contrast. methods essential machine because they directly influence behaviors have major impact models. an augmentation technique increase be reduce improve MLP combination was found four random-based search strategy. MPII datasets test proposed approach. achieved accuracy 91.2% 90.2% 87.5% 89.9% MLP. study first executed dataset.

Язык: Английский

---

Temporal–spatial transformer based motor imagery classification for BCI using independent component analysis

Biomedical Signal Processing and Control, Год журнала: 2023, Номер 87, С. 105359 - 105359

Опубликована: Авг. 25, 2023

Язык: Английский

---

Soil seismic response modeling of KiK-net downhole array sites with CNN and LSTM networks

Engineering Applications of Artificial Intelligence, Год журнала: 2023, Номер 121, С. 105990 - 105990

Опубликована: Фев. 22, 2023

Язык: Английский

---

End-to-end tire defect detection model based on transfer learning techniques

Neural Computing and Applications, Год журнала: 2024, Номер 36(20), С. 12483 - 12503

Опубликована: Апрель 22, 2024

Abstract Visual inspection of defective tires post-production is vital for human safety, as faulty can lead to explosions, accidents, and loss life. With the advancement technology, transfer learning (TL) plays an influential role in many computer vision applications, including tire defect detection problem. However, automatic difficult two reasons. The first presence complex anisotropic multi-textured rubber layers. Second, there no standard X-ray image dataset use detection. In this study, a TL-based model proposed using new from global company. First, we collected labeled consisting 3366 images 20,000 qualified tires. Although covers 15 types defects arising different design patterns, our primary focus on binary classification detect or absence defects. This challenging was split into 70, 15, 15% training, validation, testing, respectively. Then, nine common pre-trained models were fine-tuned, trained, tested dataset. These are Xception, InceptionV3, VGG16, VGG19, ResNet50, ResNet152V2, DenseNet121, InceptionResNetV2, MobileNetV2. results show that fine-tuned DenseNet21 InceptionNet achieve compatible with literature. Moreover, Xception outperformed compared TL literature methods terms recall, precision, accuracy, F1 score. it achieved testing 73.7, 88, 80.2, 94.75% score, respectively, validation 73.3, 90.24, 80.9, 95%

Язык: Английский

---