You are here

Title: Support Vector Machine based Exploratory Projection Pursuit Optimization for User face identification

Speaker: Souad Larabi Mariesainte , PhD| Information Technology Department, CCIS, KSU.


For most real-world biometric identification applications, the training database size could be very large, i.e. in the range of several thousands. This yields to the curse of dimensionality problem. The downside of such a problem is that it could negatively affect both the identification performance and speed. In this research study we used Projection Pursuit (PP) methods to determine clusters of users having significant similarities and then applied Support Vector Machine (SVM) classifiers on each cluster of users independently. Experimental results obtained using YALE face database showed that the proposed method allowed reducing the dimensionality of the dataset for training SVMs and thus improving the performance and speed of biometrics based user identification systems. Further improvement can be achieved when using within-between projection pursuit index with PP and also when using hierarchical clustering based PP.

Speaker's Bio: Dr. Souad Larabi Marie-Sainte is an assistant professor at the Department of Information Technology in the College of Computer and Information Sciences at King Saud University. Prior to joining KSU, Dr. Souad worked as an associate researcher in the department of Computer Science at Toulouse1 Capitole University, France. She also supervised undergraduate and master students at this department. She earned her PhD from the Computer Science department of Toulouse1 Capitole University, France, in the area of bio-inspired algorithms in June 2011. She holds two MSc degrees; the first one in Mathematics, Computing, Decision and Organization at Paris Dauphine University, France and the second in Computing, Mathematics and Applications at Sorbonne Paris1 University, France. Dr. Souad graduated from USTHB University in Algeria in the area of Operational Research. Her research interests include Combinatorial Optimization, Heuristics, Meta-heuristics, Artificial Intelligence and especially Bio-inspired algorithms applied to multidisciplinary domains, Bioinformatics and Algorithms Analysis and Design. She has written some papers and attended various specialized international conferences.


Date/Time: Tuesday, March 22, 2016 at 12:00pm

Location: Khadija Auditorium, F49 in Building 6  - Broadcast to Room 2090 in CCIS Building 31