Enhancing auto-tracking feature in video-based 2-dimensional motion capture
A precise and low error rate auto-tracking system in motion capture is very essential since it can save time for researcher to analyze a movement and give accurate results. Accurate data about human movements may assist sport and medical biomechanics. The purpose of present project is to create a robust auto-tracking algorithm for 2 dimensional movements by using computer program. The calculation of range of motion may illustrate the rate of algorithm accuracy. Videos containing human movements were made. The subject performed certain movement continuously and parallel to the camera-plane. The subject was attached with markers with certain colors in certain position. The algorithm means to extract the marker movements as representation of subject movements. The result has shown a good accuracy in range of motion calculation. The angle obtained from the calculation showed no divergence with manual calculation for each frame for the first 250 video frames (10 seconds) and the algorithm has been tested with some videos containing different human movements.
B00450 | (wh) | Available |
No other version available