Abstract

Human kinematic analysis remains one of the most challenging tasks for the researchers. In a typical
gait analysis study, measurement of various joint angles is carried out with the help of markers attached to the
human body. The human activity is captured by a video camera and the video is then analyzed using appropriate software. Locating the markers manually in each frame is cumbersome and may give erroneous results.
The problems like, motion blur, scattering, marker occlusion, marker reflection etc. are also prevalent. The
video recording settings play an important role in deciding the marker location precisely. The effects of input
parameters such as frame rate, and shutter speed, on the pixel area of the LED marker were investigated using
the Response Surface Methodology (RSM). It has been observed that higher shutter speed time has a negative
effect on pixel area of the LED marker and on the area of the bounding box. The experimentation using RSM
showed that shutter speed time of 2300.21 µsec and frame rate of 60.76 Frame per second (FPS) are required
to produce marker pixel and bounding box of desired quality.