OSCARS and BT Enhancement
10/9 Student Conference Presentation Rehearsal Seminar
Part I: OSCARS: A Framework on Applying Facial Detection and Recognition in Online Applications
by Caleb Robinson who is a Junior Computer Science undergraduate student at the University of Mississippi. He is currently working with Dr. Xue on the OSCARS project.
Abstract:The OSCARS project is the Olemiss Student Classroom Attendance Recognition System. The goals of the project are to create a system that uses facial detection and recognition algorithms to automate the taking of attendances in classrooms. The challenges that this presents are: the recognition of individuals based on a single image per sample, and the creation of a platform to implement the system. We have shown that realtime performance is possible using a client/server model with OpenCV by performance benchmarking. The facial detection for OSCARS is done by Local Binary Pattern Cascades and the recognition is done by Uniform Threshold LBPH.
Part II: Enhancing BitTorrent with Twitter-like Social Relationships
by Zeyang Su who is a Master Student in The Department of Computer Science here at the University of Mississippi, working with Dr. Wang. He has also earned a bachelor Degree at China University of Geosciences (Beijing).
Abstract: BitTorrent (BT) is one of the most popular file sharing protocols over the Internet. In current BT system, tit-for-tat is a well adopted mechanism to seek sharing efficiency, as such an approach is generally considered robust to peers’ selfish behaviors. Recently, more and more studies suggested that if long-term relationships among BT peers can be built and peers can cooperate with each other, the system can achieve better sharing efficiency. This intuition is based on the fact that a robust sharing incentive can be reasonably expected among friends; the tit-for-tat mechanism is thus no longer necessary across these peers.