Robust online tracking via sparse gradient convolution networks

作者：

Highlights：

• An appearance model integrates sparse representation with convolutional networks.

• The proposed methods can exploit the local spatial and structural features.

• An SGC network without offline training is proposed for object tracking.

摘要

•An appearance model integrates sparse representation with convolutional networks.•The proposed methods can exploit the local spatial and structural features.•An SGC network without offline training is proposed for object tracking.

论文关键词：Online tracking,Convolutional networks,Gradient features,Sparse representation

论文评审过程：Received 8 October 2019, Revised 18 October 2020, Accepted 30 October 2020, Available online 9 November 2020, Version of Record 12 November 2020.

论文官网地址：https://doi.org/10.1016/j.image.2020.116056