Introduction: Microsoft Kinect™ (MK) is a markerless, cheap and suitable technology that could be used to replace more complex video analysis instruments (Bonnechère et al., 2014). Due to high price and poor transportability of marker-based system, MK can be a useful instrument for sport scientists to analyse athletes’ movement. Thus, the aim of this study was to investigate the concurrent validity of the MK compared to a low cost video analysis software such as Kinovea (Ki) in the measure of knee joint angles (KJA) during a slow squat motion. Methods: Thirty-seven males took part in this study (28.29±13.56 yrs, 75.05±8.92 kg, and 180.37±8.21 cm). Participants were asked to perform three repetitions of squat exercises at three different KJA (i.e., quarter squat (QS), half squat (HS), and full squat (FS)). The KJA was measured during the three squat exercises by MK using LabVIEW (National Instrument, USA) and Ki (V0.8.15). In order to detect differences and interaction among squat exercises and between MK and Ki, two-way ANOVA RM corrected for Greenhouse-Geisser adjustment was performed. The magnitude of the difference was evaluated by partial eta squared (part 2). Pearson Correlation Coefficient (r) was computed to evaluate the relationship between MK and Ki in each squat exercise. The agreement of the two methods was assessed using a modified Bland-Altman analysis (B-A) corrected for repeated measurements. The systematic bias (mean absolute difference, calculated as Ki-KN) and limits of agreement [LOA (1.96 SD of the bias)] were calculated. Results: The two-way ANOVA RM showed significant interaction among KJA in the squat exercises and between MK and Ki (F(1.541,55.528)=16.053, p<0.001, part2=0.308). A significant correlation between MK and Ki was found in each squat exercise (QS: r=0.742, p<0.001; HS: r=0.714, p<0.001; FS: r=0.832, p<0.001). A systematic bias of 28.24 [8.04 - 48.45] degrees was detected by B-A. Discussion: While the MK ability to track a human figure is acceptable for videogame, MK measurement accuracy is not acceptable for athletes’ analysis (Pfister et al., 2014). The systematic bias found between KJA may be explained by different anatomical landmarks estimated by MK (i.e., anterior superior iliac crest) and used in Ki (i.e., greater trochanters) as affirmed by Schmitz et al. (2015) that compare the MK to a marker-based system. References: Bonnechère B, Jansen B, Salvia P, Bouzahouene H, Omelina L, Moiseev F, … Van Sint Jan S. (2014). Gait Posture, 39(1), 593–598. Pfister A, West AM, Bronner S, Noah JA. (2014). J Med Eng Technol, 38(5), 274–280. Schmitz A, Ye M, Boggess G, Shapiro R, Yang R, Noehren B. (2015). Gait Posture, 41(2), 694–698.

Validation of the microsoft kinect™ for evaluating knee joint angle in squat exercise

D. Formenti;
2016-01-01

Abstract

Introduction: Microsoft Kinect™ (MK) is a markerless, cheap and suitable technology that could be used to replace more complex video analysis instruments (Bonnechère et al., 2014). Due to high price and poor transportability of marker-based system, MK can be a useful instrument for sport scientists to analyse athletes’ movement. Thus, the aim of this study was to investigate the concurrent validity of the MK compared to a low cost video analysis software such as Kinovea (Ki) in the measure of knee joint angles (KJA) during a slow squat motion. Methods: Thirty-seven males took part in this study (28.29±13.56 yrs, 75.05±8.92 kg, and 180.37±8.21 cm). Participants were asked to perform three repetitions of squat exercises at three different KJA (i.e., quarter squat (QS), half squat (HS), and full squat (FS)). The KJA was measured during the three squat exercises by MK using LabVIEW (National Instrument, USA) and Ki (V0.8.15). In order to detect differences and interaction among squat exercises and between MK and Ki, two-way ANOVA RM corrected for Greenhouse-Geisser adjustment was performed. The magnitude of the difference was evaluated by partial eta squared (part 2). Pearson Correlation Coefficient (r) was computed to evaluate the relationship between MK and Ki in each squat exercise. The agreement of the two methods was assessed using a modified Bland-Altman analysis (B-A) corrected for repeated measurements. The systematic bias (mean absolute difference, calculated as Ki-KN) and limits of agreement [LOA (1.96 SD of the bias)] were calculated. Results: The two-way ANOVA RM showed significant interaction among KJA in the squat exercises and between MK and Ki (F(1.541,55.528)=16.053, p<0.001, part2=0.308). A significant correlation between MK and Ki was found in each squat exercise (QS: r=0.742, p<0.001; HS: r=0.714, p<0.001; FS: r=0.832, p<0.001). A systematic bias of 28.24 [8.04 - 48.45] degrees was detected by B-A. Discussion: While the MK ability to track a human figure is acceptable for videogame, MK measurement accuracy is not acceptable for athletes’ analysis (Pfister et al., 2014). The systematic bias found between KJA may be explained by different anatomical landmarks estimated by MK (i.e., anterior superior iliac crest) and used in Ki (i.e., greater trochanters) as affirmed by Schmitz et al. (2015) that compare the MK to a marker-based system. References: Bonnechère B, Jansen B, Salvia P, Bouzahouene H, Omelina L, Moiseev F, … Van Sint Jan S. (2014). Gait Posture, 39(1), 593–598. Pfister A, West AM, Bronner S, Noah JA. (2014). J Med Eng Technol, 38(5), 274–280. Schmitz A, Ye M, Boggess G, Shapiro R, Yang R, Noehren B. (2015). Gait Posture, 41(2), 694–698.
2016
Donghi, F.; Rossi, A.; Bonfanti, L.; Formenti, D.; Alberti, G.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11383/2085387
 Attenzione

L'Ateneo sottopone a validazione solo i file PDF allegati

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact