An online learned CRF model for multi-target tracking

By Bo Yang, Ramkant Nevatia. 2012 IEEE Conference on Computer Vision and Pattern Recognition, 2012

341 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/64485202ef61431497a0353dcee3d1da4b34d84f

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 Computer Vision and Pattern Recognition

Citations in the corpus

Unsupervised Hard Example Mining from Videos for Improved Object Detection (2018, background): “Spatial label dependencies are widely modeled by Markov random fields [18] and conditional random fields [53], while the smoothness of labels across time is a staple of tracking methods and other video processing algorithms [50,28,59].”
Quasi-Dense Similarity Learning for Multiple Object Tracking (2020, methodology): “Appearance similarity in MOT To exploit instance appearance similarity to strengthen tracking and re-identify vanished objects, some methods directly use an independent model [2,20,22,29,38,40,45,49] or add an extra embedding head to the detector for end-to-end training [26,44,50,53].”
PoseTrack: A Benchmark for Human Pose Estimation and Tracking (2017, methodology): “Given the joint localization threshold for each person, we compute two sets of evaluation metrics, one which is commonly used for evaluating multi-person pose estimation [35], and one from the multi-target tracking literature [50, 8, 30] to evaluate multiperson pose tracking.”
Multiple Hypothesis Tracking Revisited (2015, background): “In tracklet association, several works [2, 42, 21, 22] train discriminative appearance models of tracklets in order to design a better affinity score function.”
On pairwise costs for network flow multi-object tracking (2014, methodology): “Despite recent progress in the field [3, 5, 8, 20, 21, 22, 27], tracking remains a challenging problem especially in crowded and cluttered scenes.”
Tracking the Untrackable: Learning to Track Multiple Cues with Long-Term Dependencies (2017, background): “There are many cases that linear motion models can not deal with long-term occlusions; in this cases, non-linear motion models are proposed to produce a more accurate prediction [72, 73, 10].”
Learning by Tracking: Siamese CNN for Robust Target Association (2016, methodology): “In [57], a CRF model is learned to better distinguish pedestrians with similar appearance.”
GMMCP tracker: Globally optimal Generalized Maximum Multi Clique problem for multiple object tracking (2015, methodology): “It is worth to mention that tracklets have been previously used as reliable inputs in many tracking algorithm [26, 11, 13, 22].”
Multiple object tracking: A literature review (2014, methodology): “Each node in the graph represents observations [105] or pairs of tracklets [59], and a label is predicted to indicate which track the observations belongs to or whether to link the tracklets.”
Tracking Without Bells and Whistles (2019, methodology): “In [62], a CRF model is learned to better distinguish pedestrians with similar appearance.”
Simple online and realtime tracking with a deep association metric (2017, background): “For example, flow network formulations [1, 2, 3] and probabilistic graphical models [4, 5, 6, 7] have become popular frameworks of this type.”
Learning deep features for multiple object tracking by using a multi-task learning strategy (2014, neutral): “In contrast to another line of multiple object tracking using detection association [3, 4, 5, 6, 7], model-free multiple object tracking is an extension of single object tracking, which does not depend on class-specific detector and can online predict locations of arbitrary objects.”
Robust Online Multi-object Tracking Based on Tracklet Confidence and Online Discriminative Appearance Learning (2014, methodology): “Compared to batch tracking systems [2, 14, 20, 21], our system (p4) is still competitive.”