Accurate detection and classification of different physical exercise postures play a crucial role in monitoring fitness levels, preventing injuries, and personalizing workout routines. Traditional approaches using handcrafted feature extraction and shallow classifiers often suffer from low generalization and limited scalability. To address these limitations, this paper explores advanced deep learning models such as convolutional neural networks (CNN), recurrent neural networks (RNN), capsule networks (CapsNet), gated recurrent units (GRU), and multilayer perceptron (MLP), along with hybrid architectures, to accurately classify exercise quality categories. The models were trained on sensor data collected from smart devices, capturing motion dynamics and postural changes. Among the evaluated models, Capsule Network achieved the highest accuracy of 0.99, followed by hybrid MLP with CNN and transformer model with 0.97, demonstrating superior capability in recognizing complex activity patterns. The results show that deep learning models can effectively identify different exercise postures with high precision and recall, paving the way for intelligent fitness monitoring systems. Future work includes optimizing the models for real-time applications and extending the system to include a wider range of physical activities.
A deep learning–based HAR framework utilizing hip-mounted accelerometer and gyroscope signals from the USC-HAD dataset, which contains readings from healthy participants only, is evaluated, providing a more realistic assessment of subject-independent generalization across unseen individuals.
F. Naveed, Hamza Khan, Zaki Uddin et al.· Scientific Reports· 0 citations
Mobile health has become a popular option for patients to monitor and analyze their body activities and vital signs using their mobile devices, such as smartphones and smartwatches. In addition, the healthcare community has begun using artificial intelligence (AI) models to automate the diagnosis of abnormal conditions and diseases, particularly in real‐time emergency scenarios. This article introduces a new deep learning model to automatically identify daily body activities. We propose a hybrid AI model that combines a convolutional neural network (CNN) and a gated recurrent unit (GRU) for robust human activity classification. Key contributions include (1) the design of an end‐to‐end CNN‐GRU model that jointly extracts local spatiotemporal features and models long‐term temporal dependencies from raw sensor data, and (2) a comprehensive benchmarking framework validating the developed model against previous machine learning and deep learning classifiers. A public mobile health dataset (MHEALTH) has been used in this study. This dataset includes 12 physical activities, for example, knee bending, walking, and running. Key findings demonstrate that the CNN‐GRU model achieves a state‐of‐the‐art classification accuracy of 99.50% on the publicly available MHEALTH dataset, which encompasses 12 distinct physical activities. It significantly outperforms CNN‐LSTM (98.83%), 1‐D CNN (96.89%), and traditional ensemble methods while maintaining high precision, recall, and F1‐scores across all activity classes. Therefore, our developed model can be implemented in a cloud computing system to monitor senior patients as a critical healthcare application.
Raed Alotaibi, O. Reyad, M. E. Karar· International Journal of Tel...· 0 citations
Wearable sensor-based human activity recognition (HAR) has become increasingly popular for applications in health monitoring, fitness, and smart living. But the use of deep learning models on edge devices is still challenging due to limited memory and computational power. In this paper, we develop a resource-constrained CNN-GRU hybrid model for HAR on the WISDM dataset. This architecture uses convolutional layers for spatial learning and gated recurrent units (GRU) for sequence learning. For deployment on the edge, the model is quantized to TensorFlow Lite (TFLite) using float16. Our experiments show that the model achieves an accuracy of 94.07%, while the size of the model is substantially smaller and suitable for deployment on edge devices. Importantly, the TFLite model maintains the same accuracy as the original model, ensuring its suitability for real-time deployment. The extensive assessment through confusion matrices, ROC curves and classification metrics confirms the effectiveness of the model across various activities. The proposed approach offers a balance between accuracy and computational efficiency, enabling real-time HAR on edge devices.
Unknown authors· International journal of com...· 0 citations
With the increasing demand for intelligent sports training, traditional motion recognition methods based on visual or sensor data face challenges such as strong environmental dependence, high computational complexity, and insufficient interpretability. This paper proposes an innovative deep neural decision forest model, which builds an end-to-end motion analysis system by fusing the feature extraction ability of convolutional neural network and the hierarchical classification mechanism of decision forest. The model uses multi-granularity scanning technology to capture multi-scale features of time series data, the differentiable decision tree is used to realize the hierarchical classification of rule interpretation, and the decision path and feature importance weight are visualized to provide the coaches with “white box” decision-making basis. The experimental results show that the decision path consistency of this model reaches 94.6%, which is significantly better than the black box CNN model. The experimental results show that on the three public data sets of WISDM, UCI-HAR and DAPHNet, the model action recognition accuracy rate reaches 98.5%, and the F1 score is 98.3%, which is 4.3 percentage points higher than that of traditional convolutional neural networks. In a noisy environment (signal-to-noise ratio of 10 dB), the recognition accuracy is still 92.7%, and the inference time of embedded devices is only 12.1 ms. This study confirms the effectiveness of deep neural decision forest in complex sports scenarios. Its innovation is reflected in transforming non-differentiable decision trees into differentiable forms to achieve end-to-end training, developing multi-granularity feature extraction mechanisms to enhance time series modeling capabilities, establishing an adaptive integration framework to improve system stability, and providing a new technical path and theoretical support for precise sports training. The core innovations of this paper can be condensed into three points: (1) structure differentiability: this paper transforms the non-differentiable decision tree into differentiable form through soft binning function and Kronecker product for the first time, and realizes the end-to-end joint optimization of CNN feature extraction and decision forest classification; (2) time series multi-granularity: this paper proposes a sliding window multi-granularity scanning mechanism to synchronously capture the transient impact and long-term rhythm of motion data; (3) dynamic integration: this paper designs an adaptive weight allocation strategy based on the accuracy of validation set, which replaces the static voting mechanism of traditional random forest, and significantly improves the robustness in noisy environment. The system can not only output action labels in real time, but also generate action quality scores (0–100 points) based on the predicted probability entropy value, and automatically correlate the key feature deviations in the decision path. In the field test, the system pushes the visual feedback report to the coaches through the tablet computer with a delay of 12.1ms, which supports the formulation of personalized correction scheme, and realized the closed loop from ‘action recognition’ to ‘intelligent coaching’.
An auxiliary action-recognition evaluation framework incorporating a Big Generative Adversarial Network (BigGAN)-based data augmentation mechanism that offers a reproducible foundation for data augmentation, action classification, and intelligent feedback in sports motion monitoring applications is developed.
Fangge Zhang, Tianli Hao, Longyu He· Journal of Mechanics in Medi...· 0 citations
A hybrid deep learning framework combining convolutional neural networks and gradient boosting decision trees was studied. This framework aims to utilize multimodal physiological data to explain muscle fatigue during high-intensity interval training. In a controlled laboratory environment, participants' skin temperature, heart rate, and surface electromyography data were collected while they trained under a standardized training protocol. By using advanced denoising techniques based on empirical mode decomposition and wavelets to process the raw physiological data, participant-specific normalization and comprehensive feature extraction were performed, including spectral, temporal, and nonlinear parameters. The decision tree ensemble achieved reliable final classification and regression, while the convolutional neural network components completed hierarchical feature abstraction. By integrating Shapley additive explanations, the model becomes more transparent, making the impact of each physiological feature on fatigue progression clear. Thru systematic validation using cross-validation and independent test cohorts, the proposed framework consistently demonstrates stronger predictive and generalization capabilities compared to traditional machine learning and conventional deep learning methods. Feature contribution analysis indicates that the most important factors for predicting subject fatigue are derived from electromyography, heart rate variability, and thermal response. In order to enhance the safety and performance of athletes, this framework allows for real-time, user-friendly monitoring of high-intensity training. This work provides a detailed and interpretable approach to using data-driven sports analysis to predict muscle fatigue.
Yang Fei, Jiying Wei, Dianli Ji· International Conference on...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
