A motion model based on recurrent neural networks for visual object tracking

作者：

Highlights：

• Developing an LSTM-based motion model for single-object tracking.

• End-to-end training using different combinations of processed motion features.

• Incorporating the developed model into existing single-object tracking algorithms.

• Dynamically setting search region location.

• Demonstrating considerable improvement in tracking performance.

摘要

•Developing an LSTM-based motion model for single-object tracking.•End-to-end training using different combinations of processed motion features.•Incorporating the developed model into existing single-object tracking algorithms.•Dynamically setting search region location.•Demonstrating considerable improvement in tracking performance.

论文关键词：Single-object tracking,Motion model,Long short-term memory,Recurrent neural network

论文评审过程：Received 15 June 2022, Accepted 7 August 2022, Available online 14 August 2022, Version of Record 24 August 2022.

论文官网地址：https://doi.org/10.1016/j.imavis.2022.104533