Multi-View Real-Time Human Motion Recognition Based on Ensemble Learning

Chen, Pengyun; Wang, Xiang; Wang, Mingyang; Yang, Xiaqing; Guo, Shisheng; Jiang, Chaoshu; Cui, Guolong; Kong, Lingjiang

doi:10.1109/jsen.2021.3094548

Cited by 17 publications

(9 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Both, Dice and Jaccard, are the most drastic sample‐based metrics as seen in the results from the considered dataset. Also, the correlation index or MCC accounts for imbalances in the dataset and it is widely used in the medical domain [34]. ■ For evaluating continuous sequences of activities : The prior metrics suffer by evaluating continuous sequences of activities with unconstrained and seamless transitions.…”

Section: Resultsmentioning

confidence: 99%

“…The IoU metric is an alternative metric accounting for detected block ratio and the block length differences. The IoU is the most extreme evaluation metric for our dataset since it is a product of the Jaccard index (hard metric on its own) multiplied with a block detection term [34]. Thus the activity ‘standing up from falling’ degrades to 41.1% (Jaccard index: 41.6%), and the ‘translation’ activity to 82.4% (Jaccard index: 84.7%) as the best class.…”

Section: Resultsmentioning

confidence: 99%

“…Equation ( 28) penalises interrupted sequences even if the predictions are broadly correct and aligned with the ground truth [34].…”

Section: Intersection-over-union (Iou)mentioning

confidence: 99%

See 2 more Smart Citations

Distributed radar fusion and recurrent networks for classification of continuous human activities

Guendel

Fioranelli

Yarovoy

2022

IET Radar Sonar & Navi

View full text Add to dashboard Cite

Continuous Human Activity Recognition (HAR) in arbitrary directions is investigated in this paper using a network of five spatially distributed pulsed Ultra‐Wideband radars. While activities performed continuously and in unconstrained trajectories provide a more realistic and natural scenario for HAR, the network of radar sensors is proposed to address the issue of unfavourable or occluded perspectives when using only a single sensor. Different techniques to combine the relevant information from the multiple radars in the network are investigated, focussing on signal level fusion directly applied on Range‐Time maps, and the selection of radar nodes based on location and velocity of the target derived from multilateration processing and tracking. Recurrent Neural Networks with and without bidirectionality are used to classify the activities based on the micro‐Doppler (μD) spectrograms obtained for sensor fusion techniques. To assess classification performances, novel evaluation metrics accounting for the continuous nature of the sequence of activities and inherent imbalances in the dataset are proposed and compared with existing metrics. It is shown that the conventional accuracy metric may not capture all the important aspects for continuous HAR, and the proposed metrics can be considered for a more comprehensive evaluation.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Distributed radar fusion and recurrent networks for classification of continuous human activities

Guendel

Fioranelli

Yarovoy

2022

IET Radar Sonar & Navi

View full text Add to dashboard Cite

show abstract

“…However, the efficiency is not significantly improved, and the final local correlation needs large amounts of calculation. For example, Chen et al ( 2021b ) proposed a bottom–up method to associate some detection candidates with a single human body. However, the final detection time was relatively long.…”

Section: Literature Surveymentioning

confidence: 99%

Applying Deep Learning-Based Human Motion Recognition System in Sports Competition

Zhang

2022

Front. Neurorobot.

View full text Add to dashboard Cite

The exploration here intends to compensate for the traditional human motion recognition (HMR) systems' poor performance on large-scale datasets and micromotions. To this end, improvement is designed for the HMR in sports competition based on the deep learning (DL) algorithm. First, the background and research status of HMR are introduced. Then, a new HMR algorithm is proposed based on kernel extreme learning machine (KELM) multidimensional feature fusion (MFF). Afterward, a simulation experiment is designed to evaluate the performance of the proposed KELM-MFF-based HMR algorithm. The results showed that the recognition rate of the proposed KELM-MFF-based HMR is higher than other algorithms. The recognition rate at 10 video frame sampling points is ranked from high to low: the proposed KELM-MFF-based HMR, support vector machine (SVM)-MFF-based HMR, convolutional neural network (CNN) + optical flow (CNN-T)-based HMR, improved dense trajectory (IDT)-based HMR, converse3D (C3D)-based HMR, and CNN-based HMR. Meanwhile, the feature recognition rate of the proposed KELM-MFF-based HMR for the color dimension is higher than the time dimension, by up to 24%. Besides, the proposed KELM-MFF-based HMR algorithm's recognition rate is 92.4% under early feature fusion and 92.1% under late feature fusion, higher than 91.8 and 90.5% of the SVM-MFF-based HMR. Finally, the proposed KELM-MFF-based HMR algorithm takes 30 and 15 s for training and testing. Therefore, the algorithm designed here can be used to deal with large-scale datasets and capture and recognize micromotions. The research content provides a reference for applying extreme learning machine algorithms in sports competitions.

show abstract

“…Previous work on the classification of naturalistic sequences of human motion observed by a radar network has explored learned representations of sequences by applications of recurrent neural networks [9]. Other work utilizes logistic regression to discriminate between motion classes and defines heuristics for sensor fusion [10].…”

Section: Introductionmentioning

confidence: 99%

A Bayesian Network for the Classification of Human Motion as Observed by Distributed Radar

Svenningsson¹,

Fioranelli

Yarovoy

et al. 2022

IEEE Trans. Aerosp. Electron. Syst.

View full text Add to dashboard Cite

In this article, a statistical model of human motion as observed by a network of radar sensors is presented where knowledge on the position and heading of the target provides information on the observation conditions of each sensor node. Sequences of motions are estimated from measurements of instantaneous Doppler frequency, which captures informative micromotions exhibited by the human target. A closed-form Bayesian estimation algorithm is presented that jointly estimates the state of the target and its exhibited motion class which are described by a hidden Markov model. To correct errors in the estimated motion class distribution introduced by faulty modeling assumptions, calibration of the probability distribution and measurement likelihood is performed by isotonic regression. It is shown, by Manuscript

show abstract

Multi-View Real-Time Human Motion Recognition Based on Ensemble Learning

Cited by 17 publications

References 39 publications

Distributed radar fusion and recurrent networks for classification of continuous human activities

Distributed radar fusion and recurrent networks for classification of continuous human activities

Applying Deep Learning-Based Human Motion Recognition System in Sports Competition

A Bayesian Network for the Classification of Human Motion as Observed by Distributed Radar

Contact Info

Product

Resources

About