Criterion validity of neural networks to assess lower limb motion during cycling

Bini, Rodrigo Rico; Serrancolí, Gil; Santiago, Paulo Roberto Pereira; Pinto, Allan; Moura, Filipe A.

doi:10.1080/02640414.2023.2194725

Cited by 3 publications

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

Validity of neural networks in determining lower limb kinematics in stationary cycling

2023

View full text Add to dashboard Cite

Purpose Increasing access to marker-less technology has enabled practitioners to obtain kinematic data more quickly. However, the validation of many of these methods is lacking. Therefore, the validity of pre-trained neural networks was explored in this study compared to reflective marker tracking from sagittal plane cycling motion. Methods Twenty-six cyclists were assessed during stationary cycling at self-selected cadence and moderate intensity exercise. Standard video from their sagittal plane was obtained to extract joint kinematics. Hip, knee, and ankle angles were calculated from marker digitisation and from two deep learning-based approaches (TransPose and MediaPipe). Results Typical errors ranged between 1 and 10° for TransPose and 3–9° for MediaPipe. Correlations between joint angles calculated from TransPose and marker digitalization were stronger (0.47–0.98) than those from MediaPipe (0.25–0.96). Conclusion TransPose seemed to perform better than MediaPipe but both methods presented poor performance when tracking the foot and ankle. This seems to be associated with the low frame rate and image resolution when using standard video mode.

show abstract

Validity of neural networks in determining lower limb kinematics in stationary cycling

2023

View full text Add to dashboard Cite

show abstract

Describing posture on the bike: Do different methods lead to similar

Bini,

Lanferdini,

Diefenthaeler

2024

Preprint

View full text Add to dashboard Cite

Purpose Standardising methods to calculate joint angles is essential to enable the reproducibility of movement analysis in cycling. This study compared three methods for determining lower limb posture on the bike across three positions on the saddle. Methods Fourteen non-cyclists were assessed in two laboratory visits. The first involved determining their maximum aerobic capacity which was used in the second visit to define a sub-maximal cycling exercise intensity. Lower limb kinematics were obtained and angles for the hip, knee, and ankle were calculated using three methods (6 o’clock position, Minimum Knee Angle, and the Largest Leg Extension). Results Angles obtained at the 6 o’clock position were larger than those at the minimum knee angle and the largest leg extension for the hip (p < 0.01), knee (p < 0.01), and ankle joints (p < 0.01). Knee flexion was greater at the anterior position than the posterior (p < 0.01) and the reference (p < 0.01), with larger angles for the reference than the posterior (p < 0.01). The ankle was more dorsiflexed at the anterior vs. posterior positions (p < 0.01), anterior vs. reference positions (p < 0.01), and references vs. posterior positions (p < 0.01). Conclusion All three methods were sensitive to detect changes in saddle position but data should not be interchanged due to differences in angles between methods.

show abstract

A Machine Learning Approach for Predicting Pedaling Force Profile in Cycling

Ahmadi,

Rasoulian,

Veisari

et al. 2024

Sensors

View full text Add to dashboard Cite

Accurate measurement of pedaling kinetics and kinematics is vital for optimizing rehabilitation, exercise training, and understanding musculoskeletal biomechanics. Pedal reaction force, the main external force in cycling, is essential for musculoskeletal modeling and closely correlates with lower-limb muscle activity and joint reaction forces. However, sensor instrumentation like 3-axis pedal force sensors is costly and requires extensive postprocessing. Recent advancements in machine learning (ML), particularly neural network (NN) models, provide promising solutions for kinetic analyses. In this study, an NN model was developed to predict radial and mediolateral forces, providing a low-cost solution to study pedaling biomechanics with stationary cycling ergometers. Fifteen healthy individuals performed a 2 min pedaling task at two different self-selected (58 ± 5 RPM) and higher (72 ± 7 RPM) cadences. Pedal forces were recorded using a 3-axis force system. The dataset included pedal force, crank angle, cadence, power, and participants’ weight and height. The NN model achieved an inter-subject normalized root mean square error (nRMSE) of 0.15 ± 0.02 and 0.26 ± 0.05 for radial and mediolateral forces at high cadence, respectively, and 0.20 ± 0.04 and 0.22 ± 0.04 at self-selected cadence. The NN model’s low computational time suits real-time pedal force predictions, matching the accuracy of previous ML algorithms for estimating ground reaction forces in gait.

show abstract

Criterion validity of neural networks to assess lower limb motion during cycling

Cited by 3 publications

References 36 publications

Validity of neural networks in determining lower limb kinematics in stationary cycling

Validity of neural networks in determining lower limb kinematics in stationary cycling

Describing posture on the bike: Do different methods lead to similar

A Machine Learning Approach for Predicting Pedaling Force Profile in Cycling

Contact Info

Product

Resources

About