Robust visual tracking via online multiple instance learning with Fisher information
作者：
Highlights：
• An MIL tracker via Fisher information criterion is proposed that selects much more effective features.
• An efficient approach is proposed to optimize our Fisher information criterion objective function.
• Our tracker yields much stable and faster results (5 times in MATLAB) than MIL tracker (implemented in C++).
摘要
Highlights•An MIL tracker via Fisher information criterion is proposed that selects much more effective features.•An efficient approach is proposed to optimize our Fisher information criterion objective function.•Our tracker yields much stable and faster results (5 times in MATLAB) than MIL tracker (implemented in C++).
论文关键词：Tracking by detection,Multiple instance learning,Fisher information
论文评审过程：Received 19 October 2013, Revised 15 July 2014, Accepted 4 June 2015, Available online 18 June 2015, Version of Record 19 August 2015.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.06.004
