Feature Analysis of Smart Shoe Sensors for Classification of Gait Patterns

Sunarya, Unang; Hariyani, Yuli Sun; Cho, Taeheum; Roh, Jongryun; Hyeong, Joonho; Sohn, Illsoo; Kim, Sayup; Park, Cheolsoo

doi:10.3390/s20216253

Cited by 22 publications

(8 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Secondly, in order to validate the performance of the algorithm, seven additional features were added to the first scenario. The seven additional features consist of cress-factor, kurtosis, energy, skewness, spectral frequency, entropy, and zero-crossing, as defined Sunarya et al [23]. In the last scenario, feature selection of information gain was utilized to find the three best significant features then the features were applied to the algorithm.…”

Section: Resultsmentioning

confidence: 99%

Electrocardiogram feature selection and performance improvement of sleep stages classification using grid search

2022

Self Cite

View full text Add to dashboard Cite

Sleep analysis is often used to identify sleep-related human health. In many cases, sleep disorders could cause a particular disease. One of the approaches to detect sleep disorders is by investigating human sleep stages. However, the selection of the proper electrocardiogram (ECG) features is still considered challenging and becomes an issue to achieve the performance of the algorithm used. Therefore, it is necessary to investigate which ECG features are very significant to the performance of the algorithm. In this study, the support vector machine (SVM) method has been utilized to classify sleep stages into two classes namely awake and sleep. In order to improve the classification performances, an optimization method of grid search was used to find the best parameters of the SVM. Feature selection of information gain was then used to find the most significant ECG features. To validate the performance results, one leave-subject out cross-validation has been conducted during the implementation. There were ten subjects involved in this implementation. The ECG signals from those ten subjects were used to differentiate awake from sleep state. Based on the results, our method obtained an average accuracy of 85.46% a precision of 84.05% and a recall of 85.44% respectively.

show abstract

Section: Resultsmentioning

confidence: 99%

Electrocardiogram feature selection and performance improvement of sleep stages classification using grid search

2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…As one of the most popular machine learning algorithms for regression and classification tasks [ 32 ], the random forest (RF) algorithm is often used to determine the importance of eigengenes [ 33 ]. To further assess the relative importance of the eight PR-DE-ERSGs used to build the model in glioma, we ran the RF algorithm.…”

Section: Methodsmentioning

confidence: 99%

“…To maintain the stability of the test results, we repeated the above analysis based on 3 sets from the TCGA database, GSE4412 and GSE43378 sets from the GEO database, and CGGA set, respectively. As one of the most popular machine learning algorithms for regression and classification tasks [32], the random forest (RF) algorithm is often used to determine the importance of eigengenes [33]. To further assess the relative importance of the eight PR-DE-ERSGs used to build the model in glioma, we ran the RF algorithm.…”

Section: Model's Performance and Stabilitymentioning

confidence: 99%

A Composite Bioinformatic Analysis to Explore Endoplasmic Reticulum Stress-Related Prognostic Marker and Potential Pathogenic Mechanisms in Glioma by Integrating Multiomics Data

Fan

Nie

Zhang

et al. 2022

Journal of Oncology

View full text Add to dashboard Cite

In recent years, abnormal endoplasmic reticulum stress (ERS) response, as an important regulator of immunity, may play a vital role in the occurrence, development, and treatment of glioma. Weighted correlation network analysis (WGCNA) based on six glioma datasets was used to screen eight prognostic-related differentially expressed ERS-related genes (PR-DE-ERSGs) and to construct a prognostic model. BMP2 and HEY2 were identified as protective factors (HR < 1), and NUP107, DRAM1, F2R, PXDN, RNF19A, and SCG5 were identified as risk factors for glioma (HR > 1). QRT-PCR further supported significantly higher DRAM1 and lower SCG5 relative mRNA expression in gliomas. Our model has demonstrated excellent performance in predicting the prognosis of glioma patients from numerous datasets. In addition, the model shows good stability in multiple tests. Our model also shows broad clinical promise in predicting drug treatment effects. More immune cells/processes in the high-risk population with poor prognosis illustrate the importance of the tumor immunosuppressive environment in glioma. The potential role of the HEY2-based competitive endogenous RNA (ceRNA) regulatory network in glioma was validated and revealed the possible important role of glycolysis in glioma ERS. IDH1 and TP53 mutations with better prognosis were strongly associated with the risk score and PR-DE-ERSGs expression in the model. mDNAsi was also closely related to the risk score and clinical characteristics.

show abstract

“…Fibre-optic based pressure sensing systems are less prone to hysteresis; however, these can be easily damaged when walking [65]. As an alternative, some researchers have utilised accelerometers [50], [61], [63], and inertial measurement units [49], [62] for gait analysis. The majority of these devices are large/bulky and have not been seamlessly integrated into the wearable garment, adversely impacting the comfort of the wearer and impeding movement [61]- [63].…”

Section: Introductionmentioning

confidence: 99%

“…The use of machine learning to automate processing and analysis of large volumes of gait data has become more common in recent years. Researchers have classified gait abnormalities using shallow machine learning tools such as random forest, K nearest neighbour and support vector machine learning tools [42], [50], deep neural networks such as long short term memory networks (LSTM) [53], [69] and convolutional neural networks (CNN) [70]. The use of these tools to identify gait features has typically been more successful with multi-sensor and even multi-modality data collection [42].…”

Section: Introductionmentioning

confidence: 99%

Classifying Gait Alterations Using an Instrumented Smart Sock and Deep Learning

et al. 2022

View full text Add to dashboard Cite

This paper presents a non-invasive method of classifying gait patterns associated with various movement disorders and/or neurological conditions, utilising unobtrusive, instrumented socks and a deep learning network. Seamless instrumented socks were fabricated using three accelerometer embedded yarns, positioned at the toe (hallux), above the heel and on the lateral malleolus. Human trials were conducted on 12 able-bodied participants, an instrumented sock was worn on each foot. Participants were asked to complete seven trials consisting of their typical gait and six different gait types that mimicked the typical movement patterns associated with various movement disorders and neurological conditions. Four Neural Networks and an SVM were tested to ascertain the most effective method of automatic data classification. The Bi-LSTM generated the most accurate results and illustrates that the use of three accelerometers per foot increased classification accuracy compared to a single accelerometer per foot by 11.4%. When only a single accelerometer was utilised for classification, the ankle accelerometer generated the most accurate results in comparison to the other two. The network was able to correctly classify five different gait types: stomp (100%), shuffle (66.8%), diplegic (66.6%), hemiplegic (66.6%) and "normal walking" (58.0%). The network was incapable of correctly differentiating foot slap (21.2%) and steppage gait (4.8%). This work demonstrates that instrumented textile socks incorporating three accelerometer yarns were capable of generating sufficient data to allow a neural network to distinguish between specific gait patterns. This may enable clinicians and therapists to remotely classify gait alterations and observe changes in gait during rehabilitation.

show abstract

Feature Analysis of Smart Shoe Sensors for Classification of Gait Patterns

Cited by 22 publications

References 52 publications

Electrocardiogram feature selection and performance improvement of sleep stages classification using grid search

Electrocardiogram feature selection and performance improvement of sleep stages classification using grid search

A Composite Bioinformatic Analysis to Explore Endoplasmic Reticulum Stress-Related Prognostic Marker and Potential Pathogenic Mechanisms in Glioma by Integrating Multiomics Data

Classifying Gait Alterations Using an Instrumented Smart Sock and Deep Learning

Contact Info

Product

Resources

About