Discriminative Scale Space Tracking

By Martin Danelljan, Gustav Häger, Fahad Shahbaz Khan, Michael Felsberg. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016

1,217 citations recorded by the source corpus.

Accurate scale estimation of a target is a challenging research problem in visual object tracking. Most state-of-the-art methods employ an exhaustive scale search to estimate the target size. The exhaustive search strategy is computationally expensive and struggles when encountered with large scale variations. This paper investigates the problem of accurate and robust scale estimation in a tracking-by-detection framework. We propose a novel scale adaptive tracking approach by learning separate discriminative correlation filters for translation and scale estimation. The explicit scale filter is learned online using the target appearance sampled at a set of different scales. Contrary to standard approaches, our method directly learns the appearance change induced by variations in the target scale. Additionally, we investigate strategies to reduce the computational cost of our approach. Extensive experiments are performed on the OTB and the VOT2014 datasets. Compared to the standard exhaustive scale search, our approach achieves a gain of 2.5 percent in average overlap precision on the OTB dataset. Additionally, our method is computationally efficient, operating at a 50 percent higher frame rate compared to the exhaustive scale search. Our method obtains the top rank in performance by outperforming 19 state-of-the-art trackers on OTB and 37 state-of-the-art trackers on VOT2014.

Subfields: Artificial Intelligence, Biomedical Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/ce8c76bfedc5d86faabf0d49dc42a4924f75876d

How researchers see this paper

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

In Artificial Intelligence

In Biomedical Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Learning Spatial-Temporal Regularized Correlation Filters for Visual Tracking (2018, methodology): “Along with the introduction of feature representation [14, 28], nonlinear kernel [19], scale estimation [11, 23, 24], max-margin classifiers [43], spatial regularization [13, 18], and continuous convolution [15], DCF-based trackers have been greatly improved and significantly advanced the state-of-t”
Discriminative Correlation Filter Tracker with Channel and Spatial Reliability (2016, background): “The use of multiple channels in correlation filters (Henriques et al. 2015; Danelljan et al. 2017; Galoogahi et al. 2013) has become very popular in visual tracking.”
AutoTrack: Towards High-Performance Visual Tracking for UAV With Automatic Spatio-Temporal Regularization (2020, background): “There are currently two main research interests in this area: discriminative correlation filter (DCF)-based methods [5, 6, 7, 8, 9, 10, 11, 12, 13, 14] as well as deep learningbased approaches [15, 16, 17, 18, 19, 20].”
GOT-10k: A Large High-Diversity Benchmark for Generic Object Tracking in the Wild (2018, neutral): “227 5@CPU D Y N HOG ICCV’15 fDSST [43] 0.”
Learning Adaptive Discriminative Correlation Filters via Temporal Consistency Preserving Spatial Feature Selection for Robust Visual Object Tracking (2018, methodology): “On the other hand, fDSST [45] proposed to perform scale detection in the tracking stage.”
The Seventh Visual Object Tracking VOT2019 Challenge Results (2019, methodology): “To further accelerate tracking performance, M2C2F enhances the baseline fDSST approach [18] by exploiting a faster scale estimation method in which the target scale filter is learned jointly via sparsely sampled scalespaces.”
SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking (2019, methodology): “The introduction of correlation filter methods [7, 28, 17, 9, 36, 34] make the tracking reach an unprecedented height both in efficiency and accuracy.”
Is First Person Vision Challenging for Object Tracking? (2020, neutral): “Specifically, we have considered short-term trackers [46] based on correlation-filters with hand-crafted features (MOSSE [11], DSST [23], KCF [37], Staple [7], BACF [42], DCFNet [79], STRCF [49], MCCTH [78]) and deep features (ECO [21], ATOM [22], DiMP [9], PrDiMP [24], KYS [10]).”
Kernel Cross-Correlator (2017, methodology): “Discriminative scale space tracking (DSST) (Danelljan et al. 2017) is proposed to learn separate MOSSE on different scales, enabling estimation of both translation and scale variants, which is at the cost of repeated calculations of MOSSE.”
Unveiling the Power of Deep Tracking (2018, methodology): “We compare our method with the following state-of-the-art methods: CCOT [7], ECO [1], MDNet [13], HDT [20], FCNT [12], SRDCF [18], DeepSRDCF [40], BACF [41], Staple [3], MEEM [4], HCF [8], SAMF [42], DSST [43], MUSTER [25], Struck [44], and ASLA [45].”
ATOM: Accurate Tracking by Overlap Maximization (2018, background): “Even finding a one-parameter scale factor has turned out a formidable challenge [5, 20] and most approaches resort to the bruteforce multi-scale detection strategy with its obvious computational impact.”
Real-Time MDNet (2018, methodology): “For comparison, we employ several state-of-the-art trackers including ECO [3], MDNet [1], MDNet+IEL, SRDCF [31], C-COT [4], and top performing real-time trackers, ECO-HC [3], BACF [11], PTAV [12], CFNet [21], SiamFC [20] and DSST [32].”
Probabilistic Regression for Visual Tracking (2020, background): “Examples of the latter include visual tracking [2, 11, 22], object detection [24, 26, 35] and human pose estimation [6, 42, 48].”
Learning Aberrance Repressed Correlation Filters for Real-Time UAV Tracking (2019, methodology): “The results are compared with 20 state-of-the-art trackers with both hand-crafted based trackers and deep-based trackers, i.e., KCF [13], DSST [8], SAMF [18], MUSTER [14], BACF [15], SRDCF [7], STAPLE CA [21], MCCT-H [27], STRCF [16], ECO-HC (with gray-scale) [5], ECO [5], CCOT [9], HCF [19], ADNet ”
LaSOT: A High-Quality Benchmark for Large-Scale Single Object Tracking (2018, background): “, ECO HC [7], DSST [8], CN [11], CSK [19], KCF [20], fDSST [9], SAMF [33], SCT4 [6], STC [57] and Staple [3]) or deep features (e.”