Upper Body Posture Recognition Using Inertial Sensors and Recurrent Neural Networks

Inadequate sitting posture can cause imbalanced loading on the spine and result in abnormal spinal pressure, which serves as the main risk factor contributing to irreversible and chronic spinal deformity. Therefore, sitting posture recognition is important for understanding people’s sitting behaviors and for correcting inadequate postures. Recently, wearable devices embedded with microelectromechanical systems (MEMs) sensors, such as inertial measurement units (IMUs), have received increased attention in human activity recognition. In this study, a wearable device embedded with IMUs and a machine learning algorithm were developed to classify seven static sitting postures: upright, slump, lean, right and left bending, and right and left twisting. Four 9-axis IMUs were uniformly distributed between thoracic and lumbar regions (T1-L5) and aligned on a sagittal plane to acquire kinematic information about subjects’ backs during static-dynamic alternating motions. Time-domain features served as inputs to a signal-based classification model that was developed using long short-term memory-based recurrent neural network (LSTM-RNN) architecture, and the model’s classification performance was used to evaluate the relevance between sensor signals and sitting postures. Overall results from performance evaluation tests indicate that this IMU-based measurement and LSTM-RNN structural scheme was appropriate for sitting posture recognition.

Download Full-text

AI-powered Posture Training: Application of Machine Learning in Sitting Posture Recognition Using the LifeChair Smart Cushion

Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2019/805 ◽

2019 ◽

Author(s):

Katia Bourahmoune ◽

Toshiyuki Amagasa

Keyword(s):

Machine Learning ◽

Learning Algorithm ◽

Well Being ◽

Recognition System ◽

Training System ◽

Sensor Data ◽

Physical And Mental Health ◽

Sitting Posture ◽

Posture Monitoring ◽

Posture Recognition

Humans spend on average more than half of their day sitting down. The ill-effects of poor sitting posture and prolonged sitting on physical and mental health have been extensively studied, and solutions for curbing this sedentary epidemic have received special attention in recent years. With the recent advances in sensing technologies and Artificial Intelligence (AI), sitting posture monitoring and correction is one of the key problems to address for enhancing human well-being using AI. We present the application of a sitting posture training smart cushion called LifeChair that combines a novel pressure sensing technology, a smartphone app interface and machine learning (ML) for real-time sitting posture recognition and seated stretching guidance. We present our experimental design for sitting posture and stretch pose data collection using our posture training system. We achieved an accuracy of 98.93% in detecting more than 13 different sitting postures using a fast and robust supervised learning algorithm. We also establish the importance of taking into account the divergence in user body mass index in posture monitoring. Additionally, we present the first ML-based human stretch pose recognition system for pressure sensor data and show its performance in classifying six common chair-bound stretches.

Download Full-text

Appliance Classification by Power Signal Analysis Based on Multi-Feature Combination Multi-Layer LSTM

Energies ◽

10.3390/en12142804 ◽

2019 ◽

Vol 12 (14) ◽

pp. 2804 ◽

Cited By ~ 11

Author(s):

Jin-Gyeom Kim ◽

Bowon Lee

Keyword(s):

Short Term Memory ◽

Supply And Demand ◽

Demand Management ◽

Audio Signal ◽

Single Layer ◽

Critical Factor ◽

Classification Performance ◽

Classification Model ◽

Feature Combination ◽

Load Monitoring

The imbalance of power supply and demand is an important problem to solve in power industry and Non Intrusive Load Monitoring (NILM) is one of the representative technologies for power demand management. The most critical factor to the NILM is the performance of the classifier among the last steps of the overall NILM operation, and therefore improving the performance of the NILM classifier is an important issue. This paper proposes a new architecture based on the RNN to overcome the limitations of existing classification algorithms and to improve the performance of the NILM classifier. The proposed model, called Multi-Feature Combination Multi-Layer Long Short-Term Memory (MFC-ML-LSTM), adapts various feature extraction techniques that are commonly used for audio signal processing to power signals. It uses Multi-Feature Combination (MFC) for generating the modified input data for improving the classification performance and adopts Multi-Layer LSTM (ML-LSTM) network as the classification model for further improvements. Experimental results show that the proposed method achieves the accuracy and the F1-score for appliance classification with the ranges of 95–100% and 84–100% that are superior to the existing methods based on the Gated Recurrent Unit (GRU) or a single-layer LSTM.

Download Full-text

Improved Self-Organizing Map-Based Unsupervised Learning Algorithm for Sitting Posture Recognition System

Sensors ◽

10.3390/s21186246 ◽

2021 ◽

Vol 21 (18) ◽

pp. 6246

Author(s):

Wenyu Cai ◽

Dongyang Zhao ◽

Meiyan Zhang ◽

Yinan Xu ◽

Zhu Li

Keyword(s):

Learning Algorithm ◽

Back Propagation ◽

Recognition System ◽

Back Propagation Neural Network ◽

Recognition Algorithm ◽

Self Organizing Map ◽

Sitting Posture ◽

Long Time ◽

Posture Recognition ◽

Self Organizing

As the intensity of work increases, many of us sit for long hours while working in the office. It is not easy to sit properly at work all the time and sitting for a long time with wrong postures may cause a series of health problems as time goes by. In addition, monitoring the sitting posture of patients with spinal disease would be beneficial for their recovery. Accordingly, this paper designs and implements a sitting posture recognition system from a flexible array pressure sensor, which is used to acquire pressure distribution map of sitting hips in a real-time manner. Moreover, an improved self-organizing map-based classification algorithm for six kinds of sitting posture recognition is proposed to identify whether the current sitting posture is appropriate. The extensive experimental results verify that the performance of ISOM-based sitting posture recognition algorithm (ISOM-SPR) in short outperforms that of four kinds of traditional algorithms including decision tree-based (DT), K-means-based (KM), back propagation neural network-based (BP), self-organizing map-based (SOM) sitting posture recognition algorithms. Finally, it is proven that the proposed system based on ISOM-SPR algorithm has good robustness and high accuracy.

Download Full-text

Intelligent Brushing Monitoring Using a Smart Toothbrush with Recurrent Probabilistic Neural Network

Sensors ◽

10.3390/s21041238 ◽

2021 ◽

Vol 21 (4) ◽

pp. 1238

Author(s):

Ching-Han Chen ◽

Chien-Chun Wang ◽

Yan-Zhen Chen

Keyword(s):

Neural Network ◽

Short Term Memory ◽

Inertial Sensors ◽

Recognition Accuracy ◽

Probabilistic Neural Network ◽

Time Signal ◽

High Tech ◽

Posture Recognition ◽

High Recognition Accuracy ◽

Computational Resources

Smart toothbrushes equipped with inertial sensors are emerging as high-tech oral health products in personalized health care. The real-time signal processing of nine-axis inertial sensing and toothbrush posture recognition requires high computational resources. This paper proposes a recurrent probabilistic neural network (RPNN) for toothbrush posture recognition that demonstrates the advantages of low computational resources as a requirement, along with high recognition accuracy and efficiency. The RPNN model is trained for toothbrush posture recognition and brushing position and then monitors the correctness and integrity of the Bass Brushing Technique. Compared to conventional deep learning models, the recognition accuracy of RPNN is 99.08% in our experiments, which is 16.2% higher than that of the Convolutional Neural Network (CNN) and 21.21% higher than the Long Short-Term Memory (LSTM) model. The model we used can greatly reduce the computing power of hardware devices, and thus, our system can be used directly on smartphones.

Download Full-text

A Multimodal Music Emotion Classification Method Based on Multifeature Combined Network Classifier

Mathematical Problems in Engineering ◽

10.1155/2020/4606027 ◽

2020 ◽

Vol 2020 ◽

pp. 1-11

Author(s):

Changfeng Chen ◽

Qiang Li

Keyword(s):

Neural Networks ◽

Feature Extraction ◽

Classification Accuracy ◽

Short Term Memory ◽

Classification Performance ◽

Feature Representation ◽

Classification Model ◽

Emotion Classification ◽

Fusion Methods ◽

Music Audio

Aiming at the shortcomings of single network classification model, this paper applies CNN-LSTM (convolutional neural networks-long short-term memory) combined network in the field of music emotion classification and proposes a multifeature combined network classifier based on CNN-LSTM which combines 2D (two-dimensional) feature input through CNN-LSTM and 1D (single-dimensional) feature input through DNN (deep neural networks) to make up for the deficiencies of original single feature models. The model uses multiple convolution kernels in CNN for 2D feature extraction, BiLSTM (bidirectional LSTM) for serialization processing and is used, respectively, for audio and lyrics single-modal emotion classification output. In the audio feature extraction, music audio is finely divided and the human voice is separated to obtain pure background sound clips; the spectrogram and LLDs (Low Level Descriptors) are extracted therefrom. In the lyrics feature extraction, the chi-squared test vector and word embedding extracted by Word2vec are, respectively, used as the feature representation of the lyrics. Combining the two types of heterogeneous features selected by audio and lyrics through the classification model can improve the classification performance. In order to fuse the emotional information of the two modals of music audio and lyrics, this paper proposes a multimodal ensemble learning method based on stacking, which is different from existing feature-level and decision-level fusion methods, the method avoids information loss caused by direct dimensionality reduction, and the original features are converted into label results for fusion, effectively solving the problem of feature heterogeneity. Experiments on million song dataset show that the audio classification accuracy of the multifeature combined network classifier in this paper reaches 68%, and the lyrics classification accuracy reaches 74%. The average classification accuracy of the multimodal reaches 78%, which is significantly improved compared with the single-modal.

Download Full-text

Transformer Oil Quality Assessment Using Random Forest with Feature Engineering

Energies ◽

10.3390/en14071809 ◽

2021 ◽

Vol 14 (7) ◽

pp. 1809

Author(s):

Mohammed El Amine Senoussaoui ◽

Mostefa Brahami ◽

Issouf Fofana

Keyword(s):

Machine Learning ◽

Random Forest ◽

Oil Quality ◽

Principal Component ◽

Condition Assessment ◽

Classification Performance ◽

Transformer Oil ◽

Classification Model ◽

Insulation Degradation ◽

Transformer Oils

Machine learning is widely used as a panacea in many engineering applications including the condition assessment of power transformers. Most statistics attribute the main cause of transformer failure to insulation degradation. Thus, a new, simple, and effective machine-learning approach was proposed to monitor the condition of transformer oils based on some aging indicators. The proposed approach was used to compare the performance of two machine-learning classifiers: J48 decision tree and random forest. The service-aged transformer oils were classified into four groups: the oils that can be maintained in service, the oils that should be reconditioned or filtered, the oils that should be reclaimed, and the oils that must be discarded. From the two algorithms, random forest exhibited a better performance and high accuracy with only a small amount of data. Good performance was achieved through not only the application of the proposed algorithm but also the approach of data preprocessing. Before feeding the classification model, the available data were transformed using the simple k-means method. Subsequently, the obtained data were filtered through correlation-based feature selection (CFsSubset). The resulting features were again retransformed by conducting the principal component analysis and were passed through the CFsSubset filter. The transformation and filtration of the data improved the classification performance of the adopted algorithms, especially random forest. Another advantage of the proposed method is the decrease in the number of the datasets required for the condition assessment of transformer oils, which is valuable for transformer condition monitoring.

Download Full-text

Sitsen: Passive sitting posture sensing based on wireless devices

International Journal of Distributed Sensor Networks ◽

10.1177/15501477211024846 ◽

2021 ◽

Vol 17 (7) ◽

pp. 155014772110248

Author(s):

Miaoyu Li ◽

Zhuohan Jiang ◽

Yutong Liu ◽

Shuheng Chen ◽

Marcin Wozniak ◽

...

Keyword(s):

Radio Frequency ◽

Radio Frequency Identification ◽

High Performance ◽

Learning Algorithm ◽

Low Cost ◽

Recognition System ◽

Sitting Posture ◽

Average Accuracy ◽

Phase Variations ◽

Window Approach

Physical health diseases caused by wrong sitting postures are becoming increasingly serious and widespread, especially for sedentary students and workers. Existing video-based approaches and sensor-based approaches can achieve high accuracy, while they have limitations like breaching privacy and relying on specific sensor devices. In this work, we propose Sitsen, a non-contact wireless-based sitting posture recognition system, just using radio frequency signals alone, which neither compromises the privacy nor requires using various specific sensors. We demonstrate that Sitsen can successfully recognize five habitual sitting postures with just one lightweight and low-cost radio frequency identification tag. The intuition is that different postures induce different phase variations. Due to the received phase readings are corrupted by the environmental noise and hardware imperfection, we employ series of signal processing schemes to obtain clean phase readings. Using the sliding window approach to extract effective features of the measured phase sequences and employing an appropriate machine learning algorithm, Sitsen can achieve robust and high performance. Extensive experiments are conducted in an office with 10 volunteers. The result shows that our system can recognize different sitting postures with an average accuracy of 97.02%.

Download Full-text

In-Cylinder Pressure Based Engine Knock Classification Model for High-Compression Ratio, Automotive Spark-Ignition Engines Using Various Signal Decomposition Methods

Energies ◽

10.3390/en14113117 ◽

2021 ◽

Vol 14 (11) ◽

pp. 3117

Author(s):

Junghwan Kim

Keyword(s):

Learning Algorithm ◽

Wavelet Packet ◽

Decomposition Methods ◽

Classification Model ◽

Misclassification Rate ◽

Cylinder Pressure ◽

Intrinsic Mode Functions ◽

Normal Cycle ◽

Engine Knock ◽

Cycle Detection

Engine knock determination has been conducted in various ways for spark timing calibration. In the present study, a knock classification model was developed using a machine learning algorithm. Wavelet packet decomposition (WPD) and ensemble empirical mode decomposition (EEMD) were employed for the characterization of the in-cylinder pressure signals from the experimental engine. The WPD was used to calculate 255 features from seven decomposition levels. EEMD provided total 70 features from their intrinsic mode functions (IMF). The experimental engine was operated at advanced spark timings to induce knocking under various engine speeds and load conditions. Three knock intensity metrics were employed to determine that the dataset included 4158 knock cycles out of a total of 66,000 cycles. The classification model trained with 66,000 cycles achieved an accuracy of 99.26% accuracy in the knock cycle detection. The neighborhood component analysis revealed that seven features contributed significantly to the classification. The classification model retrained with the seven significant features achieved an accuracy of 99.02%. Although the misclassification rate increased in the normal cycle detection, the feature selection decreased the model size from 253 to 8.25 MB. Finally, the compact classification model achieved an accuracy of 99.95% with the second dataset obtained at the knock borderline (KBL) timings, which validates that the model is sufficient for the KBL timing determination.

Download Full-text

Multi-Path and Group-Loss-Based Network for Speech Emotion Recognition in Multi-Domain Datasets

Sensors ◽

10.3390/s21051579 ◽

2021 ◽

Vol 21 (5) ◽

pp. 1579 ◽

Cited By ~ 1

Author(s):

Kyoung Ju Noh ◽

Chi Yoon Jeong ◽

Jiyoun Lim ◽

Seungeun Chung ◽

Gague Kim ◽

...

Keyword(s):

Emotion Recognition ◽

Short Term Memory ◽

Domain Adaptation ◽

Classification Model ◽

Speech Emotion Recognition ◽

Target Domain ◽

Model Generalization ◽

Speech Database ◽

Emotion Labels ◽

Temporal Feature

Speech emotion recognition (SER) is a natural method of recognizing individual emotions in everyday life. To distribute SER models to real-world applications, some key challenges must be overcome, such as the lack of datasets tagged with emotion labels and the weak generalization of the SER model for an unseen target domain. This study proposes a multi-path and group-loss-based network (MPGLN) for SER to support multi-domain adaptation. The proposed model includes a bidirectional long short-term memory-based temporal feature generator and a transferred feature extractor from the pre-trained VGG-like audio classification model (VGGish), and it learns simultaneously based on multiple losses according to the association of emotion labels in the discrete and dimensional models. For the evaluation of the MPGLN SER as applied to multi-cultural domain datasets, the Korean Emotional Speech Database (KESD), including KESDy18 and KESDy19, is constructed, and the English-speaking Interactive Emotional Dyadic Motion Capture database (IEMOCAP) is used. The evaluation of multi-domain adaptation and domain generalization showed 3.7% and 3.5% improvements, respectively, of the F1 score when comparing the performance of MPGLN SER with a baseline SER model that uses a temporal feature generator. We show that the MPGLN SER efficiently supports multi-domain adaptation and reinforces model generalization.

Download Full-text