Real-time record sensitive background classifier (RSBC)

作者：

Highlights：

• Pixel based, real-time, classification history sensitive, background subtraction.

• Classification record of previous frames is used to update background model.

• Dynamic background can be detected using temporal classification history.

• Ghost effect is detected and removed using the neighbourhood classification record.

摘要

•Pixel based, real-time, classification history sensitive, background subtraction.•Classification record of previous frames is used to update background model.•Dynamic background can be detected using temporal classification history.•Ghost effect is detected and removed using the neighbourhood classification record.

论文关键词：Background subtraction,Computer vision,Mini-batch,Binary classification,Switch information,Neighbourhood search,Real-time analysis

论文评审过程：Received 15 September 2017, Revised 20 September 2018, Accepted 14 October 2018, Available online 23 October 2018, Version of Record 30 October 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.10.025