Abrupt motion tracking using a visual saliency embedded particle filter

作者：

Highlights：

• Our method is robust in dealing with abrupt motion tracking.

• Target model is first used to compute weights for salient features.

• We improve classical saliency model to increase target saliency.

• Our improved saliency model is integrated into particle filter tracking.

• Our tracker can recover tracking even under global illumination changes.

摘要

Highlights•Our method is robust in dealing with abrupt motion tracking.•Target model is first used to compute weights for salient features.•We improve classical saliency model to increase target saliency.•Our improved saliency model is integrated into particle filter tracking.•Our tracker can recover tracking even under global illumination changes.

论文关键词：Object tracking,Abrupt motion,Particle filter,Visual saliency,Covariance descriptor

论文评审过程：Received 31 March 2013, Revised 25 November 2013, Accepted 30 November 2013, Available online 12 December 2013.

论文官网地址：https://doi.org/10.1016/j.patcog.2013.11.028