Robust occlusion-aware part-based visual tracking with object scale adaptation

作者：

Highlights：

• A novel occlusion-aware part-based model is proposed.

• A new model update method which maintains the long-term memory of target appearance is improved.

• An integral pipeline aiming to the long-term tracking is proposed under the correlation filters.

• Satisfactory performance of our tracker is achieved on several challenging datasets.

摘要

•A novel occlusion-aware part-based model is proposed.•A new model update method which maintains the long-term memory of target appearance is improved.•An integral pipeline aiming to the long-term tracking is proposed under the correlation filters.•Satisfactory performance of our tracker is achieved on several challenging datasets.

论文关键词：Visual tracking,Correlation filters,Convolutional neural networks,Object occlusion,Online model update

论文评审过程：Received 24 December 2017, Revised 22 March 2018, Accepted 10 April 2018, Available online 11 April 2018, Version of Record 16 May 2018.

论文官网地址：https://doi.org/10.1016/j.patcog.2018.04.011