Year of Graduation
Development of a Prototype of Mobile Face Tracking System
The work is dedicated to the analysis of classical tracking methods (Lucas-Kanade, Camshift) and several recent algorithms (On-Line Boosting, MIL, TLD, KCF), which use machine-learning methods, in the problem of face tracking on mobile devices. The first chapter contains theoretical aspects of tracking algorithms. The second chapter is devoted to the description of the developed software complex, which consists of mobile and desktop face tracking systems. The final chapter focuses on conducted experiments of performance on mobile devices and tracking accuracy. Besides, it contains a description of the developed tracking method.