Development and validation of FootNet; a new kinematic algorithm to improve footstrike and toe-off detection in treadmill running

Keyphrases

• 5-fold Cross Validation 20%
• Analysis Bias 20%
• Ankle Dorsiflexion 20%
• Anteroposterior 40%
• Bland-Altman Analysis 20%
• Center of Mass 20%
• Contact Phase 40%
• Contact Time 20%
• Deep Learning Algorithm 20%
• Detection Accuracy 20%
• Detection Error 20%
• Detection Method 20%
• Distal Tibia 20%
• Event Detection 60%
• Final Model 20%
• Flexion Angle 20%
• Foot Strike 100%
• Foot Strike Pattern 20%
• Force Data 80%
• Gait Cycle 40%
• Generalisability 20%
• Golden Standard 20%
• Joint Kinematics 20%
• Kinematic Algorithm 100%
• Kinematics 40%
• Limits of Agreement 20%
• Long Short-term Memory Neural Network 20%
• Model Driven Architecture 20%
• Nonparametric 20%
• Plantar Flexion 20%
• Publicly Available 20%
• Root Mean Square Error 80%
• Running Biomechanics 20%
• Running Speed 40%
• Source Code 20%
• Step Data 20%
• Strike Angle 40%
• Time Bias 20%
• Toe-off 100%
• Treadmill Running 100%
• Vertical Velocity 20%

Engineering

• Biomechanics 25%
• Contact Time 25%
• Deep Learning Method 25%
• Dorsiflexion 25%
• Error Detection 25%
• Event Detection 75%
• Event Step 25%
• Gait Cycle 50%
• Input Feature 25%
• Joint Kinematics 25%
• Long Short-Term Memory 25%
• Recurrent Neural Network 25%
• Root-Mean-Squared Error 100%
• Running Speed 50%

Neuroscience

• Biomechanics 100%
• Recurrent Neural Network 100%
• Short-Term Memory 100%