Dual channel LSTM based multi-feature extraction in gait for diagnosis of Neurodegenerative diseases

Zhao, Aite; Qi, Lin; Dong, Junyu; Yu, Hui

doi:10.1016/j.knosys.2018.01.004

Cited by 77 publications

(63 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to examine performance of the proposed algorithm, four literatures [18,19,22,24] are selected to compare the algorithm performance. [18,19,22,24] adopted PhysioNet Gait Dynamics in Neurodegenerative disease [26] which is the same as this study. [18] employed stance and swing intervals series of left and right foot to do the two-class classifications (including classification of ALS vs. HC, HD vs. HC and PD vs. HC).…”

Section: Classification Performance Comparison To Other Literature Bamentioning

confidence: 99%

“…[18] employed stance and swing intervals series of left and right foot to do the two-class classifications (including classification of ALS vs. HC, HD vs. HC and PD vs. HC). [22] utilized two kinds of data as input, gait pattern data (stance and swing intervals series of left and right foot) and gait force data (vGRFs). Two-class classification including HC vs. ALS, HC vs. HD, HC vs. PD and NDDs vs. HC were compared using LOOCV as the cross validation.…”

Section: Classification Performance Comparison To Other Literature Bamentioning

confidence: 99%

“…Generalized linear regression analysis and support vector machine (SVM) were applied to perform nonlinear gait pattern classifications. Zhao et al implemented dual-channel long short-term memory (LSTM)-based multi-feature extraction on gait for diagnosis of NDDs [22]. They designed a dual-channel LSTM model to merge time series and force series recorded from NDDs patients for whole gait understanding.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Evaluation of Vertical Ground Reaction Forces Pattern Visualization in Neurodegenerative Diseases Identification Using Deep Learning and Recurrence Plot Image Feature Extraction

Lin

Wen

Setiawan

2020

Sensors

View full text Add to dashboard Cite

To diagnose neurodegenerative diseases (NDDs), physicians have been clinically evaluating symptoms. However, these symptoms are not very dependable—particularly in the early stages of the diseases. This study has therefore proposed a novel classification algorithm that uses a deep learning approach to classify NDDs based on the recurrence plot of gait vertical ground reaction force (vGRF) data. The irregular gait patterns of NDDs exhibited by vGRF data can indicate different variations of force patterns compared with healthy controls (HC). The classification algorithm in this study comprises three processes: a preprocessing, feature transformation and classification. In the preprocessing process, the 5-min vGRF data divided into 10-s successive time windows. In the feature transformation process, the time-domain vGRF data are modified into an image using a recurrence plot. The total recurrence plots are 1312 plots for HC (16 subjects), 1066 plots for ALS (13 patients), 1230 plots for PD (15 patients) and 1640 plots for HD (20 subjects). The principal component analysis (PCA) is used in this stage for feature enhancement. Lastly, the convolutional neural network (CNN), as a deep learning classifier, is employed in the classification process and evaluated using the leave-one-out cross-validation (LOOCV). Gait data from HC subjects and patients with amyotrophic lateral sclerosis (ALS), Huntington’s disease (HD) and Parkinson’s disease (PD) obtained from the PhysioNet Gait Dynamics in Neurodegenerative disease were used to validate the proposed algorithm. The experimental results included two-class and multiclass classifications. In the two-class classification, the results included classification of the NDD and the HC groups and classification among the NDDs. The classification accuracy for (HC vs. ALS), (HC vs. HD), (HC vs. PD), (ALS vs. PD), (ALS vs. HD), (PD vs. HD) and (NDDs vs. HC) were 100%, 98.41%, 100%, 95.95%, 100%, 97.25% and 98.91%, respectively. In the multiclass classification, a four-class gait classification among HC, ALS, PD and HD was conducted and the classification accuracy of HC, ALS, PD and HD were 98.99%, 98.32%, 97.41% and 96.74%, respectively. The proposed method can achieve high accuracy compare to the existing results, but with shorter length of input signal (Input of existing literature using the same database is 5-min gait signal, but the proposed method only needs 10-s gait signal).

show abstract

Section: Classification Performance Comparison To Other Literature Bamentioning

confidence: 99%

Section: Classification Performance Comparison To Other Literature Bamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Evaluation of Vertical Ground Reaction Forces Pattern Visualization in Neurodegenerative Diseases Identification Using Deep Learning and Recurrence Plot Image Feature Extraction

Lin

Wen

Setiawan

2020

Sensors

View full text Add to dashboard Cite

show abstract

“…It is not possible to show a correlation between the severities of disease and the analytical results. Most research articles have demonstrated the difference between the disease groups and the healthy controls [ 26 , 27 ]. Only one paper gave details of clinical manifestations.…”

Section: Discussionmentioning

confidence: 99%

Multiscale Approximate Entropy for Gait Analysis in Patients with Neurodegenerative Diseases

Liu

Lin

2019

Entropy

View full text Add to dashboard Cite

Neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS), Parkinson’s diseases (PD), and Huntington’s disease (HD) are not rare neurological diseases. They affect different neurological systems and present various characteristic gait abnormalities. We retrieved gait signals of the right and left feet from a public domain on the Physionet. There were 13 patients with ALS, 15 patients with PD, 20 patients with HD and 16 healthy controls (HC). We used multiscale approximate entropy (MAE) to analyze ground reaction force on both feet. Our study shows that MAE increases with scales in all tested subjects. The group HD has the highest MAE and group ALS has the lowest MAE. We can differentiate ALS from HC by MAE, while scale factors >10 in the left foot. There are few significant differences of MAE between the HC and HD. We found a good correlation of MAE between both feet in group ALS. In conclusion, our results indicate that MAE analysis of gait signals can be used for diagnosis and long-term assessment for ALS and probably HD. Similarity of MAE between both feet can also be a diagnostic marker for ALS.

show abstract

“…In [8], it is emphasized that with current methods, the diagnosis of Parkinson's disease tends to be subjective, for this reason a Kinect sensor [9] is used as a support for the diagnosis of Parkinson's, seeking to capture by means of this, the movement of patients, the movement of patients, extracting characteristics such as amplitude and frequency of movement, which are important for physicians to make a correct diagnosis of this disease. In [10], it is mentioned that currently for the detection and identification of neurodegenerative diseases, physicians tend to analyze the progress of patients, in this case, they focus on the detection of Parkinson's disease, Huntington and amyotrophic lateral sclerosis, implementing a type of RNN called LSTM, achieving a 95.67% accuracy in the identification of the gaits.…”

Section: Introductionmentioning

confidence: 99%

Abnormal gait detection by means of LSTM

Pachón¹,

Pinzón-Arenas²,

Moreno³

2020

IJECE

View full text Add to dashboard Cite

This article presents a system focused on the detection of three types of abnormal walk patterns caused by neurological diseases, specifically Parkinsonian gait, Hemiplegic gait, and Spastic Diplegic gait. A Kinect sensor is used to extract the Skeleton from a person during its walk, to then calculate four types of bases that generate different sequences from the 25 points of articulations that the Skeleton gives. For each type of calculated base, a recurrent neural network (RNN) is trained, specifically a Long short-term memory (LSTM). In addition, there is a graphical user interface that allows the acquisition, training, and testing of trained networks. Of the four trained networks, 98.1% accuracy is obtained with the database that was calculated with the distance of each point provided by the Skeleton to the Hip-Center point.

show abstract

Dual channel LSTM based multi-feature extraction in gait for diagnosis of Neurodegenerative diseases

Cited by 77 publications

References 23 publications

Evaluation of Vertical Ground Reaction Forces Pattern Visualization in Neurodegenerative Diseases Identification Using Deep Learning and Recurrence Plot Image Feature Extraction

Evaluation of Vertical Ground Reaction Forces Pattern Visualization in Neurodegenerative Diseases Identification Using Deep Learning and Recurrence Plot Image Feature Extraction

Multiscale Approximate Entropy for Gait Analysis in Patients with Neurodegenerative Diseases

Abnormal gait detection by means of LSTM

Contact Info

Product

Resources

About