Learning to associate: HybridBoosted multi-target tracker for crowded scene

By Yuan Li, Chang Huang, Ramkant Nevatia. 2009

706 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

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

Adaptive Sensitive Reweighting to Mitigate Bias in Fairness-aware Classification (2018, background): “1Learning training weights is hardly a new concept in machine learning, but usually weights are learned to help boost weak learner accuracy [40] and only seldom to satisfy other training objectives [16, 21, 34].”
MOT16: A Benchmark for Multi-Object Tracking (2016, background): “accuracy evaluates how many distinct errors such as missed targets, ghost trajectories, or identity switches are made. Another set of measures that is widely used in the tracking community is that of [30]. There are three widely used metrics introduced in that work: mostly tracked, mostly lost, and ”
Tracking the Untrackable: Learning to Track Multiple Cues with Long-Term Dependencies (2017, methodology): “v:1405.3531 , 2014. [8] W. Choi. Near-online multi-target tracking with aggregated local ow descriptor. In Proceedings of the IEEE International Conference on Computer Vision , pages 30293037, 2015. [9] W. Choi and S. Savarese. Multiple target tracking in world coordinate with single, minimally cali”
UA-DETRAC: A new benchmark and protocol for multi-object detection and tracking (2015, methodology): “, 2013) that the widely 14 used MOT evaluation metrics including the MOTA, MOTP or IDS 15 scores (Stiefelhagen et al., 2006; Li et al., 2009), do not fully reveal 16 how a MOT system performs.”
The H3D Dataset for Full-Surround 3D Multi-Object Detection and Tracking in Crowded Urban Scenes (2019, methodology): “In the tracking algorithm evaluation, CLEAR MOT metrics are used [24] which include Multi-Object Tracking Precision (MOTP), Multi-Object Tracking Accuracy (MOTA) and Mostly Tracked (MT), Mostly Lost (ML) as mentioned in [25].”
Context tracker: Exploring supporters and distracters in unconstrained environments (2011, methodology): “In MTT, most methods employ an object detector [21, 17] or background subtraction [29] to find possible object responses, whereas in our case, we automat-”
Multi-target tracking by learning local-to-global trajectory models (2015, background): “Please cite this article as: S. Zhang, et al., Multi-target tracking by learning local-to-global trajectory models, Pattern Recognition (2014), http://dx.doi.org/10.1016/j.patcog.2014.08.013i approaches can do better to overcome errors by detections and association.”
Improving Data Association by Joint Modeling of Pedestrian Trajectories and Groupings (2010, methodology): “The method however does not need the entire time window either, as global approaches [17, 20].”
Self-Supervised Moving Vehicle Tracking With Stereo Sound (2019, background): “, false positive (FP) and false negative (FN) as evaluation metrics [10, 25].”
Multiobject tracking as maximum weight independent set (2011, background): “, color, posture, speed, direction) [13, 1, 23, 11, 14], as well as spatiotemporal context [15], supporting evidence from neighboring tracks [9], and estimates of an occluder map [10] and 3D scene layout [4].”
HOTA: A Higher Order Metric for Evaluating Multi-object Tracking (2020, methodology): “Other metrics such as the trajectory-based metrics (Mostly-Tracked, Partially-Tracked, Mostly-Lost, Fragmentation) [80,44], and the metrics of Leichter and Krupka (False Negative Rate, False Positive Rate, Fragmentation Index, Merger Index, Mean Deviation) [43] are commonly shown as secondary metric”
Learning to Track: Online Multi-object Tracking by Decision Making (2015, methodology): “Recently, there is a trend on learning to track that advocates the concept of injecting learning capabilities to MOT [38, 25, 22, 20, 4].”
Joint Probabilistic Data Association Revisited (2015, background): “Mostly lost (ML) and mostly tracked (MT) scores [25] respectively represent how many targets are tracked for less than 20% and more than 80% of their life span based on ground truth trajectories (GT).”
DETRAC: A New Benchmark and Protocol for Multi-Object Tracking (2015, methodology): “Evaluating the performance of MOT methods is a complex issue, the analysis in [28] has pointed out that the widely used criteria in the literature [36, 27], such as the MOTA, MOTP or IDS scores all have somewhat problems.”
Recent advances and trends in visual tracking: A review (2011, background): “Yuan et al. [102] proposed a learning-based hierarchical approach of multi-targets tracking by progressively associating detection responses into the desired target trajectories.”
CityFlow: A City-Scale Benchmark for Multi-Target Multi-Camera Vehicle Tracking and Re-Identification (2019, methodology): “The CLEAR MOT [6] metrics are used in the MOTChallenge benchmark for evaluating multiple object tracking performance.”
Survey on Contemporary Remote Surveillance Systems for Public Safety (2010, background): “6 presents an image of tracklet tracking [26].”