Learn how AI can help streamline training analysis videos!
Presented by Dr. Marc Schuh, Principal Consultant at TNG Technology Consulting, this webinar looks at the techniques and tools used to analyse canoe training.
Biomechanics engineers typically screen training videos manually frame by frame, locating specific paddle positions and measure the paddle's angle relative to the waterline, take about 20 minutes per athlete.
Employing foreground-background separation for canoe detection and waterline derivation, and pose detection to identify the paddle, this neural network has been trained to recognise essential paddle positions for routine training analysis.
This approach significantly streamlines the analysis process, reducing the workload by an order of magnitude. The incorporation of biomechanical insights into the post-processing step helps further refine the AI results and enhance the accuracy.
About Dr. Marc Shuh
Prior to his start at TNG, Marc Schuh received his doctorate in physics and was a wheelchair sprinter over 400m at three Paralympic Games, in addition to being a World Champion and holding the European record.