Learning Attentions: Residual Attentional Siamese Network for High Performance Online Visual Tracking

By Qiang Wang, Zhu Teng, Junliang Xing, Jin Gao, Weiming Hu, S. Maybank. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018

494 citations recorded by the source corpus.

Offline training for object tracking has recently shown great potentials in balancing tracking accuracy and speed. However, it is still difficult to adapt an offline trained model to a target tracked online. This work presents a Residual Attentional Siamese Network (RASNet) for high performance object tracking. The RASNet model reformulates the correlation filter within a Siamese tracking framework, and introduces different kinds of the attention mechanisms to adapt the model without updating the model online. In particular, by exploiting the offline trained general attention, the target adapted residual attention, and the channel favored feature attention, the RASNet not only mitigates the over-fitting problem in deep network training, but also enhances its discriminative capacity and adaptability due to the separation of representation learning and discriminator learning. The proposed deep architecture is trained from end to end and takes full advantage of the rich spatial temporal information to achieve robust visual tracking. Experimental results on two latest benchmarks, OTB-2015 and VOT2017, show that the RASNet tracker has the state-of-the-art tracking accuracy while runs at more than 80 frames per second.

Subfields: Computer Vision and Pattern Recognition, Global and Planetary Change, Safety, Risk, Reliability and Quality

View Paper: https://www.semanticscholar.org/paper/6683442ae358ae4261fdcde0164f83dd1ccd621b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Global and Planetary Change

In Safety, Risk, Reliability and Quality

Citations in the corpus

Deformable Siamese Attention Networks for Visual Object Tracking (2020, methodology): “However, the attentions and deep features of the target template and the search image are computed separately in [35, 45], which limits the potential performance of the Siamese architecture.”
Graph Attention Tracking (2020, unknown): “Following this similarity-learning work, many researchers try to enhance the Siamese model for feature representation but still leverage the cross-correlation for information embedding [11, 12, 30, 9].”
Know Your Surroundings: Exploiting Scene Information for Object Tracking (2020, background): “In recent years, Siamese networks have received significant attention [2,32,31,44,18].”
AutoTrack: Towards High-Performance Visual Tracking for UAV With Automatic Spatio-Temporal Regularization (2020, background): “Currently, the state-of-theart deep trackers [14, 37, 38, 39, 40, 41] are mostly built on siamese neural network [15].”
Learning the Model Update for Siamese Trackers (2019, methodology): “While the original SiamFC tracker [1] does not perform any model update, more recent Siamese trackers [1, 41, 27, 50] have adopted (1) to update their templates.”
Distractor-aware Siamese Networks for Visual Object Tracking (2018, methodology): “Secondly, most Siamese trackers can not update the model [31, 8, 2, 36, 16].”
Unsupervised Deep Tracking (2019, methodology): “The state-of-the-art deep tracking methods [1, 46, 15, 55, 27, 60, 58, 54, 4, 19, 33, 34] typically use pretrained CNN models for feature extraction.”
Deep Learning for Visual Tracking: A Comprehensive Survey (2019, methodology): “task [111], [123], and iv) giving higher weights to important feature channels or score maps [88], [124], [128], [148].”
Siamese Cascaded Region Proposal Networks for Real-Time Visual Tracking (2018, background): “By formulating object tracking as a matching problem, Siamese trackers [2, 17, 19, 23, 45, 46, 51, 59] aim to learn offline a generic similarity function from a large set of videos.”
SiamRPN++: Evolution of Siamese Visual Tracking With Very Deep Networks (2018, background): “Additionally, these two branches can remain fixed during the tracking phase [42, 1, 16, 45, 25, 54] or updated online to adapt the appearance changes of the target [44, 43, 13].”
Transformer Meets Tracker: Exploiting Temporal Context for Robust Visual Tracking (2021, background): “The recent improvements upon Siamese trackers include attention mechanism [54], reinforcement learning [22, 51], target-aware model fine-tuning [30], unsupervised training [50, 52], sophisticated backbone networks [27, 61], cascaded frameworks [14, 49], and model update mechanisms [16, 17, 58, 60].”
Video Instance Segmentation Tracking With a Modified VAE Architecture (2020, methodology): “Detection-free tracking methods [5, 58, 37, 62, 36, 69] track objects given a manual initialization of a ﬁxed number of objects in the ﬁrst frame.”
Target-Aware Deep Tracking (2019, methodology): “The proposed algorithm is compared with the state-of-the-art trackers, including the correlation ﬁlters based trackers, such as SRDCF [9], Staple [1], MCPF [48], CCOT [10], ECO [7], BACF [17], DRT [38], STRCF [22], and MCCT-H [43]; the non-real-time deep trackers such as MDNet [29], CREST [35], LSAR”
Coarse2Fine: a two-stage training method for fine-grained visual classification (2019, background): “Visual attentionmodelsVisual attention learning proposed by [23,24] learns an attention module for CNNs, which assigns weights to different spatial locations in the raw image.”