Vision Transformers (ViT) for Blanket-Penetrating Sleep Posture Recognition Using a Triple Ultra-Wideband (UWB) Radar System

A smart detection method for sleep posture based on a flexible sleep monitoring belt and vital sign signals

People spend approximately one-third of their lives in sleep, but more and more people are suffering from sleep disorders. Sleep posture is closely related to sleep quality, so related detection is very significant. In our previous work, a smart flexible sleep monitoring belt with MEMS triaxial accelerometer and pressure sensor has been developed to detect the vital signs, snore events and sleep stages. However, the method for sleep posture detection has not been studied. Therefore, to achieve high performance, low cost and comfortable experience, this paper proposes a smart detection method for sleep posture based on a flexible sleep monitoring belt and vital sign signals measured by a MEMS Inertial Measurement Unit (IMU). Statistical analysis and wavelet packet transform are applied for the feature extraction of the vital sign signals. Then the algorithm of recursive feature elimination with cross-validation is introduced to further extract the key features. Besides, machine learning models with 10-fold cross validation process, such as decision tree, random forest, support vector machine, extreme gradient boosting and adaptive boosting, were adopted to recognize the sleep posture. 15 subjects were recruited to participate the experiment. Experimental results demonstrate that the detection accuracy of the random forest algorithm is the highest among the five machine learning models, which reaches 96.02 %. Therefore, the proposed sleep posture detection method based on the flexible sleep monitoring belt is feasible and effective.

A Deep Learning Method of Human Identification from Radar Signal for Daily Sleep Health Monitoring

Radar signal has been shown as a promising source for human identification. In daily home sleep-monitoring scenarios, large-scale motion features may not always be practical, and the heart motion or respiration data may not be as ideal as they are in a controlled laboratory setting. Human identification from radar sequences is still a challenging task. Furthermore, there is a need to address the open-set recognition problem for radar sequences, which has not been sufficiently studied. In this paper, we propose a deep learning-based approach for human identification using radar sequences captured during sleep in a daily home-monitoring setup. To enhance robustness, we preprocess the sequences to mitigate environmental interference before employing a deep convolution neural network for human identification. We introduce a Principal Component Space feature representation to detect unknown sequences. Our method is rigorously evaluated using both a public data set and a set of experimentally acquired radar sequences. We report a labeling accuracy of 98.2% and 96.8% on average for the two data sets, respectively, which outperforms the state-of-the-art techniques. Our method excels at accurately distinguishing unknown sequences from labeled ones, with nearly 100% detection of unknown samples and minimal misclassification of labeled samples as unknown.

Assessing Non-Specific Neck Pain through Pose Estimation from Images Based on Ensemble Learning

BACKGROUND

Mobile phones, laptops, and computers have become an indispensable part of our lives in recent years. Workers may have an incorrect posture when using a computer for a prolonged period of time. Using these products with an incorrect posture can lead to neck pain. However, there are limited data on postures in real-life situations.

METHODS

In this study, we used a common camera to record images of subjects carrying out three different tasks (a typing task, a gaming task, and a video-watching task) on a computer. Different artificial intelligence (AI)-based pose estimation approaches were applied to analyze the head's yaw, pitch, and roll and coordinate information of the eyes, nose, neck, and shoulders in the images. We used machine learning models such as random forest, XGBoost, logistic regression, and ensemble learning to build a model to predict whether a subject had neck pain by analyzing their posture when using the computer.

RESULTS

After feature selection and adjustment of the predictive models, nested cross-validation was applied to evaluate the models and fine-tune the hyperparameters. Finally, the ensemble learning approach was utilized to construct a model via bagging, which achieved a performance with 87% accuracy, 92% precision, 80.3% recall, 95.5% specificity, and an AUROC of 0.878.

CONCLUSIONS

We developed a predictive model for the identification of non-specific neck pain using 2D video images without the need for costly devices, advanced environment settings, or extra sensors. This method could provide an effective way for clinically evaluating poor posture during real-world computer usage scenarios.

Research on Diesel Engine Fault Status Identification Method Based on Synchro Squeezing S-Transform and Vision Transformer

The reliability and safety of diesel engines gradually decrease with the increase in running time, leading to frequent failures. To address the problem that it is difficult for the traditional fault status identification methods to identify diesel engine faults accurately, a diesel engine fault status identification method based on synchro squeezing S-transform (SSST) and vision transformer (ViT) is proposed. This method can effectively combine the advantages of the SSST method in processing non-linear and non-smooth signals with the powerful image classification capability of ViT. The vibration signals reflecting the diesel engine status are collected by sensors. To solve the problems of low time-frequency resolution and weak energy aggregation in traditional signal time-frequency analysis methods, the SSST method is used to convert the vibration signals into two-dimensional time-frequency maps; the ViT model is used to extract time-frequency image features for training to achieve diesel engine status assessment. Pre-set fault experiments are carried out using the diesel engine condition monitoring experimental bench, and the proposed method is compared with three traditional methods, namely, ST-ViT, SSST-2DCNN and FFT spectrum-1DCNN. The experimental results show that the overall fault status identification accuracy in the public dataset and the actual laboratory data reaches 98.31% and 95.67%, respectively, providing a new idea for diesel engine fault status identification.