Exploring Workout Repetition Counting and Validation Through Deep Learning

Ferreira, Bruno; Ferreira, Pedro M.; Pinheiro, Gil; Figueiredo, Nelson; Carvalho, F. P. S.; Menezes, Paulo; Batista, Jorge

doi:10.1007/978-3-030-50347-5_1

Cited by 4 publications

(8 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Counting through skeleton data Ferreira et al [26] and Stromback et al [28] studies are similar to the presented work. It adopted the pose estimation method and extracted features from human joint spatial allocation to classify and count the repetition of the motion.…”

Section: Counting On Signalssupporting

confidence: 88%

“…The other approach [26,10,11,27,28] uses the camera to record the movement and identify the temporal patterns in the video. Existing research usually solves a more general problem in counting the periodical movements but not necessarily human movements.…”

Section: Motion Repetition Countingmentioning

confidence: 99%

“…The second step is to develop a counting method based on the extracted features. The first sub-problem either rely on the neural network [26,10,11] or features extracted by domain knowledge in computer visions [27,10]. Counting the repetition from these features is done by (1) peak detection, (2) Counting through frequency domain, or (3) neural network methods.…”

Section: Motion Repetition Countingmentioning

confidence: 99%

“…Nonetheless, this valuable advantage did not explicitly state in the previous study. Besides, Ferreira et al [26] works have reported significant differences in the accuracy in different camera locations, and both studies require users to configure different hyperparameters for different kinds of exercise.…”

Section: Counting On Signalsmentioning

confidence: 99%

See 3 more Smart Citations

Viewpoint-Invariant Exercise Repetition Counting

Hsu

Efstratios

Tsui

et al. 2021

Preprint

View full text Add to dashboard Cite

Background Counting the repetition of human exercise and physical rehabilitation is a common task in rehabilitation and exercise training. The existing vision-based repetition counting methods less emphasize the concurrent motions in the same video. Methods This work analyzed the spectrogram of the pose estimation result to count the repetition. Besides from the public datasets. This work also collected exercise videos from 11 adults to verify the proposed method is capable for handling concurrent motion and different view angles. Results The presented method was validated on the University of Idaho Physical Rehabilitation Movements Data Set (UI-PRMD) and MM-fit dataset. The overall mean absolute error (MAE) for MM-fit was 0.06 with off-by-one Accuracy (OBOA) 0.94. As for UI-PRMD dataset, MAE was 0.06 with OBOA 0.95. We have also tested the performance in a variety of camera locations and concurrent motions with 57 skeleton time-series video with overall MAE 0.07 and OBOA 0.91. Conclusion The proposed method provides a view-angle and motion agnostic concurrent motion counting. This method can potentially use in large-scale remote rehabilitation and exercise training with only one camera.

show abstract

Section: Counting On Signalssupporting

confidence: 88%

Section: Motion Repetition Countingmentioning

confidence: 99%

Section: Motion Repetition Countingmentioning

confidence: 99%

Section: Counting On Signalsmentioning

confidence: 99%

See 2 more Smart Citations

Viewpoint-Invariant Exercise Repetition Counting

Hsu

Efstratios

Tsui

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The other approach [4,5,[20][21][22] uses the camera to record the movement and identifies the temporal patterns in the video. Existing research usually solves a more general problem in counting the periodical movements but not necessarily human movements.…”

Section: Motion Repetition Countingmentioning

confidence: 99%

Viewpoint-Invariant Exercise Repetition Counting

Hsu¹,

Zhang²,

Tsougenis³

et al. 2021

Preprint

View full text Add to dashboard Cite

Counting the repetition of human exercise and physical rehabilitation is a common task in rehabilitation and exercise training. The existing vision-based repetition counting methods less emphasize the concurrent motions in the same video. This work presents a vision-based human motion repetition counting applicable to counting concurrent motions through the skeleton location extracted from various pose estimation methods. The presented method was validated on the University of Idaho Physical Rehabilitation Movements Data Set (UI-PRMD), and MM-fit dataset. The overall mean absolute error (MAE) for mm-fit was 0.06 with off-by-one Accuracy (OBOA) 0.94. Overall MAE for UI-PRMD dataset was 0.06 with OBOA 0.95. We have also tested the performance in different kind of camera location and concurrent motions with conveniently collected video with overall MAE 0.06 and OBOA 0.88. The proposed method provides a viewangle and motion agnostic concurrent motion counting. This method can potentially used in large-scale remote rehabilitation and exercise training with only one camera.

show abstract