Multi-target tracking by continuous energy minimization

By Anton Andriyenko, Konrad Schindler. Computer Vision and Pattern Recognition, 2011

345 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/cc7b1da8d2879ddf83a0f1cdfb24311ba5d75d60

How researchers see this paper

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

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

GMCP-Tracker: Global Multi-object Tracking Using Generalized Minimum Clique Graphs (2012, methodology): “We tested the proposed method on the diverse sequences of Town Center [1], TUD-corssing [2], TUD-Stadtmitte [2], PETS2009 [3], and a new sequence called Parking Lot, all with promising results.”
Detection and Tracking of Occluded People (2014, methodology): “As a first extension we integrate our joint detector into two state-of-the-art people tracking approaches (Andriyenko and Schindler 2011; Pirsiavash et al. 2011) and evaluate its performance in the context of people tracking in crowded scenes.”
An online learned CRF model for multi-target tracking (2012, methodology): “Our ground truth includes all persons appearing in the video, and has one more person than that in [2].”
Tracking Without Bells and Whistles (2019, methodology): “For sequences with comparatively low frame rates, we apply a constant velocity assumption (CVA) for all objects as in [11, 2].”
Computer Vision for Autonomous Vehicles: Problems, Datasets and State-of-the-Art (2017, background): “For this highly non-convex problem, Andriyenko & Schindler (2011) use a heuristic energy minimization scheme with repeated jump moves to escape week minima and better explore the variable-dimensional search space.”
MOT16: A Benchmark for Multi-Object Tracking (2016, methodology): “Here, we use the basic approach [5] without explicit occlusion reasoning or appearance model.”
DETRAC: A New Benchmark and Protocol for Multi-Object Tracking (2015, methodology): “We evaluate six state-of-the-art target association algorithms based on the object detection including GOG [33], CEM [2], DCT [3], IHTLS [12], H2T [37], and CMOT [4].”
GMMCP tracker: Globally optimal Generalized Maximum Multi Clique problem for multiple object tracking (2015, background): “However, due to the complexity of their models, the proposed solutions are approximate [3, 4, 27].”
Deep Affinity Network for Multiple Object Tracking (2018, neutral): “8 CompACT [93]+CEM [96] 5.”
Multiple object tracking: A literature review (2014, background): “For example, the penalty term in [109] is inversely proportional to the distance between two detection responses with different trajectory labels.”
Multi-target tracking by learning local-to-global trajectory models (2015, background): “As they consider more global information, these association Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/pr Pattern Recognition http://dx.doi.org/10.1016/j.patcog.2014.08.013 0031-3203/Crown Copyright & 2014 Published by Elsevier Ltd.”
MOTChallenge 2015: Towards a Benchmark for Multi-Target Tracking (2015, methodology): “Here, we use the basic approach [7] without explicit occlusion reasoning or appearance model.”
Discrete-continuous optimization for multi-target tracking (2012, background): “Somewhat against the trend, [ 3] also belongs to the second group of non-recursive trackers, but relaxes all discrete variables to a completely continuous state space.”
High-Speed tracking-by-detection without using image information (2017, background): “solve these problems: [1, 2] define a continuous energy function and search for strong local minima using sophisticated minimization techniques.”
Learning People Detectors for Tracking in Crowded Scenes (2013, methodology): “Moreover, we employ a recent multi-target tracker [2], c.”
TrackFormer: Multi-Object Tracking with Transformers (2021, methodology): “Motion can be modelled for trajectory prediction [1, 25, 42] using a constant velocity assumption (CVA) [2, 9] or the social force model [25, 34, 43, 58].”
UA-DETRAC: A new benchmark and protocol for multi-object detection and tracking (2015, methodology): “For MOT evaluation, we also deﬁne three levels of di ﬃ cul-ties among the 40 testing sequences, i.e. , easy (10 sequences), medium (20 sequences), and hard (10 sequences) based on the average PR-MOTA score (deﬁned in Section 3.2) of the MOT systems constructed from combinations of six representative”