Residual LSTM Attention Network for Object Tracking

Authors:
Publisher website
Google Scholar
Abstract
In this Letter, we propose an attention network for object tracking. To construct the proposed attention network for sequential data, we combine long-short term memory (LSTM) and residual framework into residual LSTM (RLSTM). The LSTM which learns temporal correlation is used for temporal learning of object tracking. In the proposed RLSTM method, the residual framework, which achieves the highest accuracy in ILSVRC 2016, learns variations of spatial inputs and thus achieves spatio-temporal attention of the target object. Also, rule-based RLSTM learning is used for robust attention. Experimental results on large tracking benchmark datasets OTB-2013, OTB-100, and OTB-50 show that the proposed RLSTM tracker achieves the highest performance among existing trackers including Siamese trackers, attention trackers, and correlation trackers, and also has comparable performance with the state-of-the-art deep trackers.
Funding Information
  • BK21 PLUS program

This publication has 33 references indexed in Scilit: