Adaptive skew-sensitive ensembles for face recognition in video surveillance
作者：
Highlights：
• Specialized strategy to select nontarget samples based on trajectories to design databases.
• Generate base classifiers for SS ensembles on distinct imbalances/complexities in video–video FR.
• Design active SS selection/fusion function using data with distinct imbalances and complexities.
• Adapt quantification method HDx for enhanced operational imbalance estimation in video–video FR.
• Extensive empirical results for Skew-Sensitive ensembles on synthetic and video data sets.
摘要
Highlights•Specialized strategy to select nontarget samples based on trajectories to design databases.•Generate base classifiers for SS ensembles on distinct imbalances/complexities in video–video FR.•Design active SS selection/fusion function using data with distinct imbalances and complexities.•Adapt quantification method HDx for enhanced operational imbalance estimation in video–video FR.•Extensive empirical results for Skew-Sensitive ensembles on synthetic and video data sets.
论文关键词：Adaptive classifier ensembles,Boolean combination,Imbalance estimation,Video-to-video face recognition,Video surveillance,Adaptive multiple classifier systems
论文评审过程：Received 25 November 2014, Revised 29 March 2015, Accepted 8 May 2015, Available online 19 May 2015, Version of Record 16 July 2015.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.05.008
