Activity classification in persons with stroke based on frequency features

Laudanski, Annemarie; Brouwer, Brenda; Li, Qingguo

doi:10.1016/j.medengphy.2014.11.008

Cited by 35 publications

(30 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, their algorithms are usually proprietary and nondisclosed, they do not identify all of the aforementioned activities, or are not always patient and user friendly (e.g., bulky). Furthermore, to date only a few studies have validated their algorithms on patients whose movement apparatus has been affected [18] , [19] , [20] . In these patients there is a broad range of activity levels, ranging from being able to walk only very short bouts with the help of walking aids (1 st week postoperatively), to uninhibited movement at the level of a healthy individual.…”

Section: Introductionmentioning

confidence: 99%

Clinical validation of a body-fixed 3D accelerometer and algorithm for activity monitoring in orthopaedic patients

Lipperts

Laarhoven²,

Senden³

et al. 2017

Journal of Orthopaedic Translation

View full text Add to dashboard Cite

SummaryBackground/ObjectiveActivity is increasingly being recognized as a highly relevant parameter in all areas of healthcare for diagnosis, treatment, or outcome assessment, especially in orthopaedics where the movement apparatus is directly affected. Therefore, the aim of this study was to develop, describe, and clinically validate a generic activity-monitoring algorithm, satisfying a combination of three criteria. The algorithm must be able to identify, count, and time a large set of relevant daily activities. It must be validated for orthopaedic patients as well as healthy individuals, and the validation must be in a setting that mimics free-living conditions.MethodsUsing various technical solutions, such as a dual-axis approach, dynamic inclinometry (hip flexion), and semiautomatic calibration (gait speed), the algorithms were designed to count and time the following postures, transfers, and activities of daily living: resting/sitting, standing, walking, ascending and descending stairs, sit–stand transitions, and cycling. In addition, the number of steps per walking bout was determined. Validation was performed with healthy individuals and patients who had undergone unilateral total joint arthroplasty, representing a wide spectrum of functional capacity. Video observation was used as the gold standard to count and time activities in a validation protocol approaching free-living conditions.ResultsIn total 992 and 390 events (activities or postures) were recorded in the healthy group and patient group, respectively. The mean error varied between 0% and 2.8% for the healthy group and between 0% and 7.5% for the patient group. The error expressed in percentage of time varied between 2.0% and 3.0% for both groups.ConclusionActivity monitoring of orthopaedic patients by counting and timing a large set of relevant daily life events is feasible in a user- and patient-friendly way and at high clinical validity using a generic three-dimensional accelerometer and algorithms based on empirical and physical methods. The algorithms performed well for healthy individuals as well as patients recovering after total joint replacement in a challenging validation set-up. With such a simple and transparent method real-life activity parameters can be collected in orthopaedic practice for diagnostics, treatments, outcome assessment, or biofeedback.

show abstract

Section: Introductionmentioning

confidence: 99%

Clinical validation of a body-fixed 3D accelerometer and algorithm for activity monitoring in orthopaedic patients

Lipperts

Laarhoven²,

Senden³

et al. 2017

Journal of Orthopaedic Translation

View full text Add to dashboard Cite

show abstract

“…The first five FFT coefficients (as calculated over each segment) were used, as these contain the main frequency components (up to 5 Hz). The features selected by the algorithm are also used in Shoaib et al [ 1 , 7 ] and are considered useful for running on smartphones, as they have very low or medium computational and storage complexity.…”

Section: Pre-processingmentioning

confidence: 99%

“…In the first class of techniques, the signal is divided into consecutive windows of fixed length, which, in the case of physical activities (level walking and stair walking) lie in the range of 1–10 s [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 ]. One limitation of this approach is that if an activity lasts for shorter or longer time periods than the pre-defined chosen window length, the subsequent classification might be affected.…”

Section: Introductionmentioning

confidence: 99%

Pre-Processing Effect on the Accuracy of Event-Based Activity Segmentation and Classification through Inertial Sensors

Fida

Bernabucci

Bibbo

et al. 2015

Sensors

View full text Add to dashboard Cite

Inertial sensors are increasingly being used to recognize and classify physical activities in a variety of applications. For monitoring and fitness applications, it is crucial to develop methods able to segment each activity cycle, e.g., a gait cycle, so that the successive classification step may be more accurate. To increase detection accuracy, pre-processing is often used, with a concurrent increase in computational cost. In this paper, the effect of pre-processing operations on the detection and classification of locomotion activities was investigated, to check whether the presence of pre-processing significantly contributes to an increase in accuracy. The pre-processing stages evaluated in this study were inclination correction and de-noising. Level walking, step ascending, descending and running were monitored by using a shank-mounted inertial sensor. Raw and filtered segments, obtained from a modified version of a rule-based gait detection algorithm optimized for sequential processing, were processed to extract time and frequency-based features for physical activity classification through a support vector machine classifier. The proposed method accurately detected >99% gait cycles from raw data and produced >98% accuracy on these segmented gait cycles. Pre-processing did not substantially increase classification accuracy, thus highlighting the possibility of reducing the amount of pre-processing for real-time applications.

show abstract

“…Segments identified with the above-mentioned rules were taken as the reference for all the remaining inertial data. For each segment, a set of time and frequency domain features that are used in the literature for the activity recognition problem were derived from each axis and magnitude of the accelerometer and gyroscope signal [5,21]. A total number of 138 features were then extracted.…”

Section: Feature Extraction and Selectionmentioning

confidence: 99%

Real time event-based segmentation to classify locomotion activities through a single inertial sensor

Fida

Bibbo

Bernabucci

et al. 2015

Proceedings of the 5th EAI International Conference on Wireless Mobile Communication and Healthcare - "Transforming Healthcare

View full text Add to dashboard Cite

We propose an event-based dynamic segmentation technique for the classification of locomotion activities, able to detect the midswing, initial contact and end contact events. This technique is based on the use of a shank-mounted inertial sensor incorporating a tri-axial accelerometer and a tri-axial gyroscope, and it is tested on four different locomotion activities: walking, stair ascent, stair descent and running. Gyroscope data along one component are used to dynamically determine the window size for segmentation, and a number of features are then extracted from these segments. The event-based segmentation technique has been compared against three different fixed window size segmentations, in terms of classification accuracy on two different datasets, and with two different feature sets. The dynamic event-based segmentation showed an improvement in terms of accuracy of around 5% (97% vs. 92% and 92% vs. 87%) and 1-2% (89% vs. 87% and 97% vs. 96%) for the two dataset, respectively, thus confirming the need to incorporate an event-based criterion to increase performance in the classification of motion activities.

show abstract

Activity classification in persons with stroke based on frequency features

Cited by 35 publications

References 13 publications

Clinical validation of a body-fixed 3D accelerometer and algorithm for activity monitoring in orthopaedic patients

Clinical validation of a body-fixed 3D accelerometer and algorithm for activity monitoring in orthopaedic patients

Pre-Processing Effect on the Accuracy of Event-Based Activity Segmentation and Classification through Inertial Sensors

Real time event-based segmentation to classify locomotion activities through a single inertial sensor

Contact Info

Product

Resources

About