Have a personal or library account? Click to login
Can machine learning distinguish between elite and non-elite rowers? Cover

Can machine learning distinguish between elite and non-elite rowers?

International Journal of Computer Science in Sport
Volume 24 (2025): Issue 1 (February 2025)
By: Kristine Fjellkårstad Orten,  Sander Elias Magnussen Helgesen,  Bihui Chen,  Adel Baselizadeh,  Jim Torresen and  Henrik Herrebrøden  
Open Access
|May 2025

Figures & Tables

Figure 1:

A rower using the rowing ergometer with reflective markers (marked with red).
A rower using the rowing ergometer with reflective markers (marked with red).

Figure 2:

Sequence splitting. Example visualization of rowing sequence input data. The sequence of rowing strokes was divided into multiple parts of single strokes (dark red box) with a 0.25-sec error margin (light red) to ensure a complete stroke was included. Each new sequence had a 1-second overlap with the previous sequence.
Sequence splitting. Example visualization of rowing sequence input data. The sequence of rowing strokes was divided into multiple parts of single strokes (dark red box) with a 0.25-sec error margin (light red) to ensure a complete stroke was included. Each new sequence had a 1-second overlap with the previous sequence.

Figure 3:

The GRU-CNN model extended the initial GRU model with a CNN and more layers. The input (1) was sent through a convolutional layer (2) with 128 filters, a kernel size of 3 and ReLU activation, then a max pooling layer (3) with pool size 2, then two GRU layers (4, 5) with the same parameters as the original GRU model. At last, the data is sent through a dense layer (6) with 128 units and ReLU activation, then a dropout layer (7) with a dropout rate of 0.3 and then a final one-unit dense layer (8) with sigmoid activation.
The GRU-CNN model extended the initial GRU model with a CNN and more layers. The input (1) was sent through a convolutional layer (2) with 128 filters, a kernel size of 3 and ReLU activation, then a max pooling layer (3) with pool size 2, then two GRU layers (4, 5) with the same parameters as the original GRU model. At last, the data is sent through a dense layer (6) with 128 units and ReLU activation, then a dropout layer (7) with a dropout rate of 0.3 and then a final one-unit dense layer (8) with sigmoid activation.

Figure 4:

Rowing angles. Example visualization of the second type of input features: Angles between reflective markers. From the left: α, β and γ.
Rowing angles. Example visualization of the second type of input features: Angles between reflective markers. From the left: α, β and γ.

Figure 5:

Violin plot showing validation and test accuracies for different feature combinations in the GRU model. The mean value and range are marked in the plot. The mean was computed across 10 instances of training the model. The left-most plot shows the accuracies achieved for different input features when using the validation data set, while the right-most plot shows the accuracies for the test data set.
Violin plot showing validation and test accuracies for different feature combinations in the GRU model. The mean value and range are marked in the plot. The mean was computed across 10 instances of training the model. The left-most plot shows the accuracies achieved for different input features when using the validation data set, while the right-most plot shows the accuracies for the test data set.

Figure 6:

10-fold cross-validation results for the GRU-CNN model. The mean value and range are marked in the plot.
10-fold cross-validation results for the GRU-CNN model. The mean value and range are marked in the plot.

Figure 7:

Training and validation accuracies for the MLP model for coordinate input feature combinations.
Training and validation accuracies for the MLP model for coordinate input feature combinations.

Performance metrics for the best-trained GRU-CNN model when evaluated on the test dataset_

Input featuresF1 scoreAUCAccuracy
CoordsShoulders and hips0.42260.40260.5181
Shoulders and seat0.41500.44600.5181
Shoulders and ergometer front0.69890.53760.7744
Ergometer handle and front0.67920.53680.7632
Hips and ergometer front0.52220.51170.5209

Performance metrics for the best-trained MLP model when evaluated on the test dataset_

Input featuresF1 scoreAUCAccuracy
CoordsShoulders and hips0.28360.99620.6178
Shoulders and seat0.35570.99490.6591
Shoulders and ergometer front0.93280.99960.9482
Ergometer handle and front0.91900.99990.9387
Hips and ergometer front0.68120.99480.8365
All joints111
DOI: https://doi.org/10.2478/ijcss-2025-0007 | Journal eISSN: 1684-4769
Journal RSS Feed
Language: English
Page range: 118 - 132
Published on: May 1, 2025
Published by: International Association of Computer Science in Sport
In partnership with: Paradigm Publishing Services
Publication frequency: 2 issues per year
Keywords:
Sports Analysis,
Rowing,
Machine Learning,
Neural Networks,
Deep Learning,
Gated Recurrent Unit,
Convolutional Neural Network,
Mulit-Layer Perceptron
Related subjects:
Computer sciences,
Databases and data mining,
Computer sciences, other,
Sports and recreation,
Physical education,
Sports and recreation, other

© 2025 Kristine Fjellkårstad Orten, Sander Elias Magnussen Helgesen, Bihui Chen, Adel Baselizadeh, Jim Torresen, Henrik Herrebrøden, published by International Association of Computer Science in Sport
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Previous articleVolume 24 (2025): Issue 1 (February 2025)Next article