Self-Paced Learning for Long-Term Tracking

By James Steven Supancic III, Deva Ramanan. 2013 IEEE Conference on Computer Vision and Pattern Recognition, 2013

312 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/1c721511e4c0e21bd264ca71c0d909528511b7ad

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

Citations in the corpus

Transfer Learning Based Visual Tracking with Gaussian Processes Regression (2014, methodology): “, fragments voting based [1,9,5], incorporating spatial constraints between the parts [42,37], alignment-pooling across the local patches [14]; iv) and the widely followed trackingby-detection (or discriminative) methods [6,7,20,2,8,21,31,45], which treat the tracking problem as a classification tas”
A Self-Paced Multiple-Instance Learning Framework for Co-Saliency Detection (2015, methodology): “By virtue of its generality, the SPL theory has been widely applied to various tasks, such as object tracking [29], image classification [30], and multimedia event detection [15, 16].”
MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labels (2017, background): “After that, many predeﬁned curriculums were proposed, e.g. , in (Supancic & Ramanan, 2013; Jiang et al., 2014; 2015; Sangineto et al., 2016; Chang et al., 2017; Ma et al., 2017a;b).”
Self-Paced Curriculum Learning (2015, methodology): “…been successfully applied to various applications, such as action/event detection (Jiang et al. 2014b), reranking (Jiang et al. 2014a), domain adaption (Tang et al. 2012), dictionary learning (Tang, Yang, and Gao 2012), tracking (Supančič III and Ramanan 2013) and segmentation (Kumar et al. 2011).”
Learning to Track: Online Multi-object Tracking by Decision Making (2015, background): “In single object tracking, the state-of-the-art trackers [3, 15, 16, 5, 41, 39, 34, 40] focus on how to learn a strong appearance model of the target online and use it for tracking.”
Adaptive Decontamination of the Training Set: A Unified Formulation for Discriminative Visual Tracking (2016, methodology): “Continuous weights: Most existing discriminative trackers [1, 16, 25, 29] rely on binary decisions for including or removing potential training samples.”
Curriculum Learning: A Survey (2021, methodology): “NIPS 2012 [130] CV DA, detection SPL loss, domain iterative data Supancic and Ramanan CVPR 2013 [131] CV long term object tracking SPL SVM objective iterative, stages data”
Understanding and Diagnosing Visual Tracking Systems (2015, methodology): “This method was proposed and evaluated in [30].”
Transferring Rich Feature Hierarchies for Robust Visual Tracking (2015, methodology): “Also, some methods focus on long-term tracking, e.g. [17, 27].”
Self-Paced Learning with Diversity (2014, methodology): “By virtue of its generality, various applications based on the SPL have been proposed very recently [7, 8, 11, 12, 13].”
MUlti-Store Tracker (MUSTer): A cognitive psychology inspired approach to object tracking (2015, methodology): “In [44], long-term tracking was conducted based on a self-paced learning scheme in which the target appearance was conservatively learned by selecting trustworthy frames.”
MentorNet: Regularizing Very Deep Neural Networks on Corrupted Labels (2017, background): “The weighting scheme is called self-paced learning and have been applied to many vision problems [46, 26, 38, 22].”
What Objective Does Self-paced Learning Indeed Optimize? (2015, background): “and computer vision tasks, such as object detector adaptation [6], specific-class segmentation learning [7], visual category discovery [8], and long-term tracking [9].”
Staple: Complementary Learners for Real-Time Tracking (2015, background): “Note that the TLD [21] and self-paced learning [35] algorithms also incorporate some aspects that are well-suited to the long-term tracking problem.”
Co-Saliency Detection via a Self-Paced Multiple-Instance Learning Framework (2017, methodology): “For example, Supancic et al. [39] used the formalism of SPL to automatically learn robust appearance model in object tracking.”
Track and Segment: An Iterative Unsupervised Approach for Video Object Proposals (2016, methodology): “, [20, 59, 55, 48, 17]) share our goal of localizing the same object over the video, but require human-annotation as initialization in the firstframe.”
Learning to Teach (2018, background): “As a comparison, self-paced learning (SPL) (Kumar et al., 2010; Lee & Grauman, 2011; Jiang et al., 2014a;b; Supancic & Ramanan, 2013) quantiﬁes the hardness by the loss on data.”
MEEM: Robust Tracking via Multiple Experts Using Entropy Minimization (2014, methodology): “For a more complete comparison, we also include three more recent trackers in this experiment: LSHT [12], LSST [25] and SPLTT [23].”