Human Gait-labeling Uncertainty and a Hybrid Model for Gait Segmentation

Wu, Jiaen; Maurenbrecher, Henrik; Schaer, Alessandro; Becsek, Barna Errol Mario; Easthope, Chris; Chatzipirpiridis, George; Ergeneman, Olgaç; Nelson, BradleyJ.

doi:10.36227/techrxiv.16924102.v1

Cited by 3 publications

(4 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…8 In addition, the algorithm only uses three signals to classify the gait phases, all of which come from a single gyroscope only. This reduces the amount of information used to characterize the gait cycle compared to the studies that use both gyroscope and accelerometers [31], and thus saves the energy and the cost of the wearable device. Moreover, no prior calibration is required for the sensor data, which is a significant improvement compared to other studies that need additional calibration procedures [32].…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Real-Time Gait Phase Detection on Wearable Devices for Unsupervised Gait

Becsek²,

Schaer³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Detecting gait phases unobtrusively and reliably in real-time for long-term unsupervised walking is important for clinical gait rehabilitation and early diagnosis of neurological diseases. Due to hardware limitations in wearable devices, real-time gait phase detection remains a challenge for unsupervised mobility assessment. In this work, a hybrid algorithm combining a reduced support vector machine (RSVM) and a finite state machine (FSM) is developed to address this. The RSVM is developed by reducing the SVM model size for real-time applications through a novel technique based on K-means clustering, and the FSM validates predictions and corrects misclassifications of RSVM. With the proposed algorithm, the model size is reduced by 88%, and the computation time is reduced by a factor of 36, with only a minor degradation in performance. The real-time classification performance of the algorithm is evaluated by twenty healthy subjects walking outdoors with unsupervised gait. The proposed algorithm demonstrated promising real-time performance, with the accuracy, sensitivity, and specificity of 91.51%, 91.70%, 95.77% across all test subjects.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Gait phases were labeled by a rule-based detection algorithm, which has been validated with the motion capture system and shows high accuracy [31]. Most importantly, it is not constrained by the lab environment and can label outdoor walking data at any time and anywhere for real-time algorithm validation.…”

Section: Data Labelingmentioning

confidence: 99%

Real-Time Gait Phase Detection on Wearable Devices for Unsupervised Gait

Becsek²,

Schaer³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…In addition, the algorithm only uses three signals to classify the gait phases, all of which come from a single gyroscope only. This reduces the amount of information used to characterize the gait cycle compared to the studies that use both gyroscope and accelerometers [22], and thus saves the energy and the cost of the wearable device. Moreover, no prior calibration is required for the sensor data, which is a significant improvement compared to the studies that need additional calibration procedures [32].…”

Section: Discussionmentioning

confidence: 99%

“…The total number of free walking strides collected for real-time algorithm validation is around 50000. Gait phases were labeled by a rule-based detection algorithm, which has been validated with the motion capture system and shows high accuracy [22]. Most importantly, it is not constrained by the lab environment and can label outdoor walking data at any time and anywhere for real-time algorithm validation.…”

Section: Data Acquisitionmentioning

confidence: 99%

Real-Time Gait Phase Detection on Wearable Devices for Unsupervised Gait

Becsek²,

Schaer³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Detecting gait phases unobtrusively and reliablyin real-time for long-term unsupervised walking isimportant for clinical gait rehabilitation and early diagnosisof neurological diseases. Due to hardware limitations inwearable devices (e.g., memory and computation power),reliable real-time gait phase detection remains a challengefor unsupervised mobility assessment. In this work, a hybridalgorithm combining a reduced support vector machine(RSVM) and a finite state machine (FSM) is developedto address this. K-means clustering is used to reduce thenumber of support vectors (SVs) by constructing a smallerdataset that contains the most informative data points.For each gait phase prediction, an FSM is designed tovalidate the prediction and correct misclassifications. AfterSV reduction, the model size is reduced by 88%, and thecomputation time is reduced by a factor of 36, with onlya minor degradation in prediction performance of 4.12%,2.34%, and 4.85% for sensitivity, specificity, and accuracy,respectively. The real-time classification performance of thealgorithm is evaluated by twenty healthy subjects walkingalong a predefined route with unsupervised free-living gait.The proposed algorithm demonstrated promising real-timeperformance, with an accuracy of 91.51%, a sensitivity of91.70%, and a specificity of 95.77% across all test subjects.The algorithm also demonstrated its robustness with respectto different values of walking speed, cadence, andstride length.

show abstract

Classification of Center of Mass Acceleration Patterns in Older People with Knee Osteoarthritis and Fear of Falling

González-Olguín

Rodríguez

Córdoba³

et al. 2022

IJERPH

View full text Add to dashboard Cite

(1) Background: The preoccupation related to the fall, also called fear of falling (FOF) by some authors is of interest in the fields of geriatrics and gerontology because it is related to the risk of falling and subsequent morbidity of falling. This study seeks to classify the acceleration patterns of the center of mass during walking in subjects with mild and moderate knee osteoarthritis (KOA) for three levels of FOF (mild, moderate, and high). (2) Method: Center-of-mass acceleration patterns were recorded in all three planes of motion for a 30-meter walk test. A convolutional neural network (CNN) was implemented for the classification of acceleration signals based on the different levels of FOF (mild, moderate, and high) for two KOA conditions (mild and moderate). (3) Results: For the three levels of FOF to fall and regardless of the degree of KOA, a precision of 0.71 was obtained. For the classification considering the three levels of FOF and only for the mild KOA condition, a precision of 0.72 was obtained. For the classification considering the three levels of FOF and only the moderate KOA condition, a precision of 0.81 was obtained, the same as in the previous case, and finally for the classification for two levels of FOF, a high vs. moderate precision of 0.78 was obtained. For high vs. low, a precision of 0.77 was obtained, and for the moderate vs. low, a precision of 0.8 was obtained. Finally, when considering both KOA conditions, a 0.74 rating was obtained. (4) Conclusions: The classification model based on deep learning (CNN) allows for the adequate discrimination of the acceleration patterns of the moderate class above the low or high FOF.

show abstract

Human Gait-labeling Uncertainty and a Hybrid Model for Gait Segmentation

Cited by 3 publications

References 4 publications

Real-Time Gait Phase Detection on Wearable Devices for Unsupervised Gait

Real-Time Gait Phase Detection on Wearable Devices for Unsupervised Gait

Real-Time Gait Phase Detection on Wearable Devices for Unsupervised Gait

Classification of Center of Mass Acceleration Patterns in Older People with Knee Osteoarthritis and Fear of Falling

Contact Info

Product

Resources

About