Unveiling the Power of Deep Tracking

By Goutam Bhat, Joakim Johnander, Martin Danelljan, Fahad Shahbaz Khan, Michael Felsberg. European Conference on Computer Vision, 2018

438 citations recorded by the source corpus.

In the field of generic object tracking numerous attempts have been made to exploit deep features. Despite all expectations, deep trackers are yet to reach an outstanding level of performance compared to methods solely based on handcrafted features. In this paper, we investigate this key issue and propose an approach to unlock the true potential of deep features for tracking. We systematically study the characteristics of both deep and shallow features, and their relation to tracking accuracy and robustness. We identify the limited data and low spatial resolution as the main challenges, and propose strategies to counter these issues when integrating deep features for tracking. Furthermore, we propose a novel adaptive fusion approach that leverages the complementary properties of deep and shallow features to improve both robustness and accuracy. Extensive experiments are performed on four challenging datasets. On VOT2017, our approach significantly outperforms the top performing tracker from the challenge with a relative gain of 17% in EAO.

Subfields: Computer Vision and Pattern Recognition, Environmental Engineering

View Paper: https://www.semanticscholar.org/paper/0eb1b75b98d4f4f69a5fb7669ad86d85cdd76848

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 Environmental Engineering

Citations in the corpus

Energy-Based Models for Deep Probabilistic Regression (2020, background): “Dataset Metric ECO SiamFC MDNet UPDT DaSiamRPN SiamRPN++ ATOM ATOM∗ Ours [9] [1] [38] [2] [64] [31] [8]”
Learning Spatio-Temporal Transformer for Visual Tracking (2021, background): “3 shows that STARK-S50 achieves a competitive result, which is better than DiMP [3] and UPDT [4].”
Deep Learning for Visual Tracking: A Comprehensive Survey (2019, methodology): “To achieve the goal of learning generic representations for target modeling and constructing a more robust target models, the main contributions of methods are classified into: i) offline training of CNNs on large-scale datasets for visual tracking [63], [68], [80], [89], [97], [100], [101], [104], ”
Autoregressive Visual Tracking (2023, background): “Visual object tracking [5, 20, 34, 38, 48, 52] is a foundational objective in the realm of computer vision, whereby the tracker endeavors to estimate the location of an arbitrary target in each video frame, based on its initial state.”
Deeper and Wider Siamese Networks for Real-Time Visual Tracking (2019, background): “There is a recent work also studying how to leverage deep networks for visual tracking [3].”
Know Your Surroundings: Exploiting Scene Information for Object Tracking (2020, methodology): “We compare our approach with the recent state-of-the-art trackers: DRT [41], RCO [29], UPDT [4], DaSiamRPN [52], MFT [29], LADCF [48], ATOM [8], SiamRPN++ [31], and DiMP-50 [3].”
Learning Discriminative Model Prediction for Tracking (2019, methodology): “Given the first frame with annotation, we employ data augmentation strategies [3] to construct an initial set Strain containing 15 samples.”
Learning the Model Update for Siamese Trackers (2019, methodology): “We compare with trackers including DRT [37], DeepSTRCF [28], LSART [38], R MCPF [48], SRCT [25], CSRDCF [30], LADCF [44], MFT [22], UPDT [2] and ATOM [8] among others as in [22].”
Target-Aware Deep Tracking (2019, background): “Deep features from multiple layers have also integrated for visual tracking [10, 32, 7, 3].”
SiamRPN++: Evolution of Siamese Visual Tracking With Very Deep Networks (2018, methodology): “Even though, due to the lack of adaptation to the template, the robustness still has a gap with the state-of-art correlation filters methods [2] which relies on the online updating.”
The Seventh Visual Object Tracking VOT2019 Challenge Results (2019, methodology): “In contrast with UPDT, the updating interval of the correlation ﬁlters is shortened in LSRDFT.”
How to Train Your Energy-Based Model for Regression (2020, neutral): “MDNet UPDT DaSiamRPN ATOM SiamRPN++ DiMP SiamRCNN PrDiMP DiMP- DiMP- DiMP[38] [3] [58] [7] [28] [4] [51] [8] KLD-IS NCE NCE+ TrackingNet 60.”
A review of computer vision–based structural health monitoring at local and global levels (2020, background): “Currently, examples of deep learning–based visual tracking algorithms are Siamese FC,173 Siamese Mask,174 Siamese RPN+ + ,175 MFT,176 and UPDT.177 Although the deep learning– based object tracking algorithms have made great progress in accuracy and robustness, they require large volume of datasets a”
Multimodal End-to-End Autonomous Driving (2019, background): “based [10], [11], [12], [13], [14], [15], and multimodal [16], [17], [18], [19], [20], [21], [22], [23]); object tracking (imagebased [24], [25], [26], [27], [28], [29], [30], multimodal [31]); traffic sign recognition [32]; semantic segmentation (by class/category [33], [34], [35] and object instan”
Learning Target Candidate Association to Keep Track of What Not to Track (2021, methodology): “We compare with state-of-theart trackers SuperDiMP [8], PrDiMP50 [12], UPDT [3], SiamRPN++ [27], ECO [10], DiMP [1], CCOT [11], MDNet [33], ATOM [9], and DaSiamRPN [49].”