Introduction - If you have any usage issues, please Google them yourself
The proposed method is tested
on 12 challenging video sequences containing targets with
abrupt appearance variations, and demonstrates that it outperforms
current state-of-the-art methods in accuracy