Robust multiple cameras pedestrian detection with multi-view Bayesian network

作者：

Highlights：

• A multi-view Bayesian network is proposed to model pedestrian candidates and their occlusion relationships in all views.

• A parameter learning algorithm is developed for MvBN by using a set of auxiliary, real-valued, and continuous variables.

• A height-adaptive projection is proposed to make the final detection robust to synthesis noises and calibration errors.

• Our approach is recognized as the best performer in five PETS evaluations from 2009 to 2013.

摘要

Highlights•A multi-view Bayesian network is proposed to model pedestrian candidates and their occlusion relationships in all views.•A parameter learning algorithm is developed for MvBN by using a set of auxiliary, real-valued, and continuous variables.•A height-adaptive projection is proposed to make the final detection robust to synthesis noises and calibration errors.•Our approach is recognized as the best performer in five PETS evaluations from 2009 to 2013.

论文关键词：Pedestrian detection,Multiple cameras,Multi-view model,Bayesian inference,Height adaptive projection

论文评审过程：Received 28 May 2014, Revised 14 September 2014, Accepted 7 December 2014, Available online 17 December 2014.

论文官网地址：https://doi.org/10.1016/j.patcog.2014.12.004