Discipline: Technology and Engineering
Subcategory: Computer Science & Information Systems
Jordan Lafontant - Howard University
Co-Author(s): Patricia Abu and Xavier Jeaspher Eser, Ateneo de Manila University, Quezon City, Metro Manila, Philippines
The purpose of my research is to test the use of the Abu algorithm as a real time video processing algorithm. If proven it suggests that the Abu algorithm is a valid tool to process videos at real time speed. A set of data from the Agency for Science, Technology and Research’s Perceptual Computing group is used in order to gauge the results against the Compressive Sensing Mixture of Gaussian (CS MOG), a state-of-the-art technique used for background subtraction. While the CS MOG is very accurate method of background subtracting, its processing rate is slightly below the required magnitude to be considered for real-time applications , which is a framerate of 30 frames per second (fps). The Abu (A) algorithm reading in frames on average at 36.46 fps and grants an average accuracy of approximately 91.2% suggest the proposed A algorithm is a good technique to consider for applications involving real time background subtraction and real time video processing. The Abu algorithm although requiring further testing seems to be a viable option to process videos, it may be possible to improve the algorithm’s speed and accuracy under further testing as well as make it more accurate to more dynamic video.
Funder Acknowledgement(s): This study was supported, in part, by a grant from the National Science Foundation awarded to Lorraine Fleming, Wayne Patterson, and Mohamed Chouikha, Principal and Co-Principal Investigators of the Global Education, Awareness and Research Undergraduate Program (GEAR UP), Howard University, Washington, DC. This study was supported and spearheaded, in large part, by the work, mentorship, and intellectual knowledge of Patricia Angela Abu, Proceso Fernandez, and Xavier Jeaspher Eser at Ateneo de Manila University, Quezon City, Metro Manila, Philippines.
Faculty Advisor: Radscheda R Nobles, radscheda.nobles@bison.howard.edu
Role: I was responsible for testing the Abu algorithm’s performance in regards to background subtraction. I analyzed the algorithm’s performance speed and accuracy while it processed a video stream as well as compared it to the control data set.