Correction to: Human activity recognition from sensor data using spatial attention-aided CNN with genetic algorithm

Sarkar, Apu; Hossain, Sabbir; Sarkar, Ram

doi:10.1007/s00521-022-08189-y

Cited by 3 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…CNN, a widely used deep neural network, plays a vital role in predicting multivariate time series by automatically extracting high-level feature representations and capturing essential insights . Specifically, CNN’s use involves convolutional operations that establish local connectivity and global sharing across multivariate time series, effectively revealing interrelationships .…”

Section: Methodsmentioning

confidence: 99%

“…Due to its carefully designed architecture, LSTM excels in learning long-term temporal dependencies. 37 The Bi-LSTM learns its parameters through both forward and backward paths. The forward layer of the Bi-LSTM updates its parameters conventionally, while the backward layer computes the derivative of the propagation error from the forward layer.…”

Section: Encoder Module 331 Stacked Bi-lstm Networkmentioning

confidence: 99%

See 1 more Smart Citation

Hybrid Deep Learning Approach for Hydraulic Fracture Morphology Diagnosis in Coalbed Methane Wells

Yang,

Gou,

Min

et al. 2024

Energy Fuels

View full text Add to dashboard Cite

Accurately diagnosing hydraulic fracture morphology is crucial for coalbed methane (CBM) development and strategy formulation. However, achieving precise and efficient diagnosis of fracture morphology remains a challenge in complex geological and engineering environments. Therefore, this study proposes a hybrid deep learning method based on multivariate time series classification. In this method, the fracturing curve is compared to the "language" of the CBM reservoir, with fracture morphology representing the semantics of this "language". This method integrates a wavelet transform module and a deep encoder−decoder module. The wavelet transform module aims to reduce signal noise and enhance model robustness, while the deep encoder−decoder module extracts latent features from the fracturing curve. Integrating hydraulic fracture morphology recognition into a hybrid deep learning model composed of diverse deep neural networks tailored for specific predictive tasks maximizes the strengths of each network component. This approach was evaluated using real data from CBM blocks in Northern China, supported by ablation experiments, numerical simulations, and microseismic interpretation, demonstrating its effectiveness and applicability. In contrast to traditional time series deep learning models, this method enables more robust extraction of advanced temporal features and delivers high-precision predictions. This novel approach provides an effective solution for diagnosing hydraulic fracturing morphology and contributes to guiding refracturing.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Encoder Module 331 Stacked Bi-lstm Networkmentioning

confidence: 99%

Hybrid Deep Learning Approach for Hydraulic Fracture Morphology Diagnosis in Coalbed Methane Wells

Yang,

Gou,

Min

et al. 2024

Energy Fuels

View full text Add to dashboard Cite

show abstract

AReNet: Cascade learning of multibranch convolutional neural networks for human activity recognition

Boudjema,

Titouna,

Titouna

2023

Multimed Tools Appl

View full text Add to dashboard Cite

Transfer Learning Approach for Human Activity Recognition Based on Continuous Wavelet Transform

2023

View full text Add to dashboard Cite

Over the last few years, human activity recognition (HAR) has drawn increasing interest from the scientific community. This attention is mainly attributable to the proliferation of wearable sensors and the expanding role of HAR in such fields as healthcare, sports, and human activity monitoring. Convolutional neural networks (CNN) are becoming a popular approach for addressing HAR problems. However, this method requires extensive training datasets to perform adequately on new data. This paper proposes a novel deep learning model pre-trained on scalograms generated using the continuous wavelet transform (CWT). Nine popular CNN architectures and different CWT configurations were considered to select the best performing combination, resulting in the training and evaluation of more than 300 deep learning models. On the source KU-HAR dataset, the selected model achieved classification accuracy and an F1 score of 97.48% and 97.52%, respectively, which outperformed contemporary state-of-the-art works where this dataset was employed. On the target UCI-HAPT dataset, the proposed model resulted in a maximum accuracy and F1-score increase of 0.21% and 0.33%, respectively, on the whole UCI-HAPT dataset and of 2.82% and 2.89%, respectively, on the UCI-HAPT subset. It was concluded that the usage of the proposed model, particularly with frozen layers, results in improved performance, faster training, and smoother gradient descent on small HAR datasets. However, the use of the pre-trained model on sufficiently large datasets may lead to negative transfer and accuracy degradation.

show abstract

Correction to: Human activity recognition from sensor data using spatial attention-aided CNN with genetic algorithm

Cited by 3 publications

References 0 publications

Hybrid Deep Learning Approach for Hydraulic Fracture Morphology Diagnosis in Coalbed Methane Wells

Hybrid Deep Learning Approach for Hydraulic Fracture Morphology Diagnosis in Coalbed Methane Wells

AReNet: Cascade learning of multibranch convolutional neural networks for human activity recognition

Transfer Learning Approach for Human Activity Recognition Based on Continuous Wavelet Transform

Contact Info

Product

Resources

About