Skip to main content
. 2023 Mar 1;4:e6. doi: 10.1017/wtc.2022.30

Table 2.

Summary of PA detection results by using the proposed approach on one representative participant’s TA muscle during the ankle joint at different postures

Results of the PA detection
RMSE [°]
Angle 5° DBSCAN K-means HAC UltraTrack ImageJ-5% ImageJ-10% Data Size
Trial 1 0.419 1.140 0.723 1.939 4.714 6.169 20
Trial 2 0.539 1.171 0.374 2.170 3.373 4.771 26
Trial 3 0.484 1.220 1.210 2.525 3.795 4.138 24
Angle 10°
Trial 1 0.358 0.357 0.994 1.698 2.424 3.848 15
Trial 2 0.892 1.143 0.980 0.934 2.817 3.863 25
Trial 3 1.053 0.632 0.539 1.649 2.787 3.357 23
Angle 15°
Trial 1 0.647 1.134 1.800 1.892 4.077 4.278 20
Trial 2 1.118 0.489 0.538 2.729 3.848 6.327 19
Trial 3 0.523 0.300 0.494 2.115 4.553 5.671 20
Overall 0.727 0.948 0.941 1.951 3.598 4.713 192

Abbreviations: DBSCAN, density-based spatial clustering of applications with noise; HAC, hierarchical agglomerative clustering; PA, pennation angle; RMSE, root mean square error; TA, tibialis anterior.