SiamFC++: Towards Robust and Accurate Visual Tracking with Target Estimation Guidelines

By Yinda Xu, Zeyu Wang, Zuoxin Li, Yuan Ye, Gang Yu. AAAI Conference on Artificial Intelligence, 2019

961 citations recorded by the source corpus.

Visual tracking problem demands to efficiently perform robust classification and accurate target state estimation over a given target at the same time. Former methods have proposed various ways of target state estimation, yet few of them took the particularity of the visual tracking problem itself into consideration. Based on a careful analysis, we propose a set of practical guidelines of target state estimation for high-performance generic object tracker design. Following these guidelines, we design our Fully Convolutional Siamese tracker++ (SiamFC++) by introducing both classification and target state estimation branch (G1), classification score without ambiguity (G2), tracking without prior knowledge (G3), and estimation quality score (G4). Extensive analysis and ablation studies demonstrate the effectiveness of our proposed guidelines. Without bells and whistles, our SiamFC++ tracker achieves state-of-the-art performance on five challenging benchmarks(OTB2015, VOT2018, LaSOT, GOT-10k, TrackingNet), which proves both the tracking and generalization ability of the tracker. Particularly, on the large-scale TrackingNet dataset, SiamFC++ achieves a previously unseen AUC score of 75.4 while running at over 90 FPS, which is far above the real-time requirement.

Subfields: Computer Vision and Pattern Recognition, Electrical and Electronic Engineering, Environmental Engineering

View Paper: https://www.semanticscholar.org/paper/be412c7c7128cf91455233b652d6c94a6001a7c8

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

1. Huchuan Lu
2. Luc Van Gool
3. Haibin Ling
4. Ming-Hsuan Yang
5. Houwen Peng
6. Xiaogang Wang
7. Wanli Ouyang
8. Martin Danelljan

In Electrical and Electronic Engineering

1. Huchuan Lu
2. Bo Li
3. Haibin Ling
4. Xin Chen
5. Houwen Peng
6. Wanli Ouyang
7. Wei Wu
8. Dong Wang

In Environmental Engineering

1. Wanli Ouyang
2. Bo Li
3. Heng Fan
4. Haibin Ling
5. Jianlong Fu
6. Wei Wu
7. Raquel Urtasun
8. Binh Yang

Citations in the corpus

• Graph Attention Tracking (2020, neutral): “Almost all current state-of-the-art Siamese trackers like SiamRPN [16], SiamRPN++ [15], SiamFC++ [33] and SiamCAR [10] utilize a cross-correlation based layer for information embedding, which takes convolution on deep features as the basic operation.”
• SwinTrack: A Simple and Strong Baseline for Transformer Tracking (2021, methodology): “IoU-aware design has been very popular recently, but most works consider IoU prediction as an auxiliary branch to assist classification or bounding box regression [52, 2, 44].”
• LasHeR: A Large-Scale High-Diversity Benchmark for RGBT Tracking (2021, neutral): “…verify the diversity and challenge of our proposed LasHeR dataset, we present the results of some RGB trackers on LasHeR dataset (RGB videos), LaSOT [8] and TrackingNet [63] in Table VI, including TransT [66], SiamFC++ [67], PrDiMP-50 [68],PrDiMP-18 [68], DiMP-50 [45], DiMP-18 [45] and ATOM [64].”
• SeqTrack: Sequence to Sequence Learning for Visual Object Tracking (2023, background): “For the classification head network, most trackers adopt stacked convolutional layers with various loss functions, including cross-entropy loss [7, 27], focal loss [51, 55], modified ℓ 2 loss [3, 12], and KL-divergence loss [13].”
• Learning Spatio-Temporal Transformer for Visual Tracking (2021, methodology): “The tracking pipelines of previous trackers [25, 51, 60, 47] are complicated.”
• Backbone is All Your Need: A Simplified Architecture for Visual Object Tracking (2022, methodology): “The compared trackers include SiamFC [2](ECCVW16), ATOM [11] (CVPR19), SiamRPN++ [42](CVPR19), DiMP [3](ICCV19), SiamFC++ [42] (AAAI20), Ocean [49](ECCV20), SiamBAN [9](CVPR20), SiamAtt [45] (CVPR 20), TransT [38](CVPR21), TrDiMP [39](CVPR21), KeepTrack [27](ICCV21), STARKS [43](ICCV21), AutoMatch [”
• Is First Person Vision Challenging for Object Tracking? (2020, neutral): “We also considered deep siamese networks (SiamFC [8], GOTURN [36], DSLT [52], SiamRPN++ [48], SiamDW [91], UpdateNet [90], SiamFC++ [86], SiamBAN [15], Ocean [92]), trackingby-detection methods (MDNet [62], VITAL [73]), as well as trackers based on target segmentation representations (SiamMask [80],”
• Transformer Tracking (2021, methodology): “Some further improvements have been made, such as adding additional branches [42, 45], using deeper architectures [21], exploiting anchor-free architectures [44, 48], and so on.”