Learning Discriminative Model Prediction for Tracking

By Goutam Bhat, Martin Danelljan, Luc Van Gool, Radu Timofte. IEEE International Conference on Computer Vision, 2019

1,338 citations recorded by the source corpus.

The current strive towards end-to-end trainable computer vision systems imposes major challenges for the task of visual tracking. In contrast to most other vision problems, tracking requires the learning of a robust target-specific appearance model online, during the inference stage. To be end-to-end trainable, the online learning of the target model thus needs to be embedded in the tracking architecture itself. Due to the imposed challenges, the popular Siamese paradigm simply predicts a target feature template, while ignoring the background appearance information during inference. Consequently, the predicted model possesses limited target-background discriminability. We develop an end-to-end tracking architecture, capable of fully exploiting both target and background appearance information for target model prediction. Our architecture is derived from a discriminative learning loss by designing a dedicated optimization process that is capable of predicting a powerful model in only a few iterations. Furthermore, our approach is able to learn key aspects of the discriminative loss itself. The proposed tracker sets a new state-of-the-art on 6 tracking benchmarks, achieving an EAO score of 0.440 on VOT2018, while running at over 40 FPS. The code and models are available at https://github.com/visionml/pytracking.

Subfields: Computer Vision and Pattern Recognition, Global and Planetary Change

View Paper: https://www.semanticscholar.org/paper/2c8315ae713b3e27c6e9f291a158134d9c516166

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

Citations in the corpus

Joint Feature Learning and Relation Modeling for Tracking: A One-Stream Framework (2022, background): “Earlier Siamese trackers [1, 26, 50] and discriminative trackers [2, 7] belong to Fig.”
Transformer Meets Tracker: Exploiting Temporal Context for Robust Visual Tracking (2021, methodology): “In other tracking frameworks with update mechanisms [20, 39, 8, 58, 60, 3], previous prediction results are collected to incrementally update the tracking model.”
Know Your Surroundings: Exploiting Scene Information for Object Tracking (2020, methodology): “While current state-of-the-art methods [8,3,31] only rely on the target appearance model to process every frame independently, our approach also propagates information about the scene from previous frames.”
LasHeR: A Large-Scale High-Diversity Benchmark for RGBT Tracking (2021, methodology): “[10] propose multi-fusion strategy in DiMP framework [39] which bases on ResNet.”
Backbone is All Your Need: A Simplified Architecture for Visual Object Tracking (2022, background): “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 [”
CoTracker: It is Better to Track Together (2023, background): “We mask out points that have not yet been initialized in the current sliding window, which allows initializing tracks from any point in the video, including in the middle of a window, and to join windows together.”
Probabilistic Regression for Visual Tracking (2020, methodology): “We demonstrate the effectiveness of our general approach by integrating it into the recent state-of-the-art tracker DiMP [3].”
Towards More Flexible and Accurate Object Tracking with Natural Language: Algorithms and Benchmark (2021, methodology): “449 006 DiMP50 [4] ICCV-2019 ResNet-50 3 3 Local 40@GPU 0.”
AiATrack: Attention in Attention for Transformer Visual Tracking (2022, methodology): “family of approaches [11,3] optimize the network parameters during the inference.”
Segment as Points for Efficient Online Multi-Object Tracking and Segmentation (2020, methodology): “The data association can be accomplished on either the 2D image plane [4,6,7,14,33,42,38] or the 3D world space [1,8,10,20,25,39].”
Learning Spatio-Temporal Transformer for Visual Tracking (2021, methodology): “Previous Siamese trackers [28, 59, 16, 7] only exploit the spatial information for tracking, while online methods [63, 66, 11, 3] use historical predictions for model updates.”
ODTrack: Online Dense Temporal Token Learning for Visual Tracking (2024, methodology): “To incorporate temporal information into the model, some approaches commonly design online updating techniques, such as updating templates(Yan et al. 2021a; Cui et al. 2022) and updating model parameters(Bhat et al. 2019).”
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.”
Ocean: Object-aware Anchor-free Tracking (2020, methodology): “We refer readers to [2,4] for more details.”
Transforming Model Prediction for Tracking (2022, methodology): “The model predictor estimates the weights of the compact DCF target model to localize the target in the test frame.”