#TITLE_ALTERNATIVE#

• Main Author: ARLIANA (NIM 13204074), RISKA
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/11289
• Institution: Institut Teknologi Bandung

Gait analysis is a science that studies the creature locomotion. Nowadays, gait analysis has been developed and used for different purposes. In the medical field, gait analysis is used to identify the cause of movement-related problems from a patient which has certain disease by identifying gait parameters. The result of gait analysis can help doctor to choose the best therapy for patient. Besides being applied in medical field, gait analysis has been advanced rapidly in sport. Furthermore, athletes and coaches use gait analysis to improve the athlete's performance. In Indonesia, gait analysis with image processing has not been developed yet, for medical or sport purpose. Therefore, in this early development phase there is a need to advance the digital image processing system technology in gait analysis so it can be used for many purposes.<p> <br /> <br /> <br /> <br /> <br /> In this final project, we developed marker detection module for 2-D gait analysis based on digital image processing. Basically, this final project research is divided into two steps: data collecting and data processing.<p> <br /> <br /> <br /> <br /> <br /> The first step is data collecting. In this research, we took several video data of the human walking movement. These videos were captured by video camera with 30 fps. Later on, the human walking movement was analyzed by putting marker in hip, knee and ankle of the right leg. Within this scope, the setting area is controlled so the marker will have a higher intensity compare to the environment areas.<p> <br /> <br /> <br /> <br /> <br /> After collecting the data, the next step is data processing. The system will detect the marker from the video by using SIMULINK and MATLAB R2007B. The marker detection is being done by changing RGB into HSV image. Then the value channel image is thresholded to produce binary image. The result of this operation is binary image which divides high intensity area with the background. Furthermore, within that area, we did the calculation of the centroid points. At the image output, the centroid point is the marker which has been detected.<p> <br /> <br /> <br /> <br /> <br /> This method has been tested with three different lighting behaviors and it has successfully identified the marker with an optimum way under four 10 Watt ultraviolet lamps. Under UV lighting condition, this method has shown a high successful rate which is 99%. Furthermore, under 18 Watt light bulb the successful rate is only 87% and the performance is also decreased into 76% if the test method was using four 10 Watt ultraviolet lamps and 18 Watt light bulb.