GMMCP tracker: Globally optimal Generalized Maximum Multi Clique problem for multiple object tracking

By Afshin Dehghan, Shayan Modiri Assari, Mubarak Shah. Computer Vision and Pattern Recognition, 2015

307 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Environmental Engineering, Ocean Engineering

View Paper: https://www.semanticscholar.org/paper/557c8065de736743ebe0f7166752903423652719

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 Environmental Engineering

In Ocean Engineering

Citations in the corpus

See the Forest for the Trees: Joint Spatial and Temporal Recurrent Neural Networks for Video-Based Person Re-identification (2017, unknown): “The bounding boxes are generated by a DPM detector [10] and a GMMCP tracker [8].”
UA-DETRAC: A new benchmark and protocol for multi-object detection and tracking (2015, methodology): “…e-mail: slyu@albany.edu (Siwei Lyu) dinia et al., 2012; Yang and Nevatia, 2012; Huang et al., 2013; Yang et al., 2014; Wen et al., 2014; Dehghan et al., 2015), require a pre-trained detector, e.g. , (Felzenszwalb et al., 2010; Doll´ar et al., 2014; Girshick et al., 2014; Yan et al.,…”
Re-ranking Person Re-identification with k-Reciprocal Encoding (2017, methodology): “Each sequence is automatically obtained by the DPM as pedestrian detector and the GMMCP [7] as tracker.”
Multiple object tracking: A literature review (2014, methodology): “Extend dynamic programming [138], linear programming [139, 140, 141], quadratic boolean programming [142], K-shortest paths [44, 19], set cover [143] and subgraph multicut [144, 145, 146], maximum multi-clique [147] are adopted to solve the association problem among detections or tracklets.”
Person Re-identification: Past, Present and Future (2016, methodology): “Each row represents a tracklet generated by the DPM detection + GMMCP tracker [125].”
Spatial-Temporal Relation Networks for Multi-Object Tracking (2019, background): “According to what kind of information is used to establishing the association between objects in different frames, the existing methods can be categorized into online methods [2, 12, 19, 27, 33, 35, 38, 47, 50, 58, 59, 9], and offline methods [7, 14, 36, 39, 40, 44, 51, 52, 53, 62].”
Deep Affinity Network for Multiple Object Tracking (2018, methodology): “[41] perform data association using a larger number of frames.”
DETRAC: A New Benchmark and Protocol for Multi-Object Tracking (2015, background): “, [41, 6, 7, 24, 29, 39, 22, 37, 11], follow the trackingby-detection paradigm, which entails an object detection Figure 1: Precision vs.”
Features for Multi-target Multi-camera Tracking and Re-identification (2018, background): “Methods that use all pairwise terms, not only timeconsecutive ones, are significantly more accurate but NPhard [18, 25, 27, 28, 41, 58, 61, 65, 66, 67].”
Online Multi-object Tracking Using CNN-Based Single Object Tracker with Spatial-Temporal Attention Mechanism (2017, methodology): “These off-line methods consider MOT as a global optimization problem and focus on designing various optimization algorithm such as network flow [39, 58], continuous energy minimization [33], max weight independent set [6], k-partite graph [56, 10], subgraph multi-cut [43, 44] and so on.”
Mask-Guided Contrastive Attention Model for Person Re-identification (2018, methodology): “There are 20,478 video sequences and 1,191,003 bounding boxes which are generated by a DPM detector [16] and a GMMCP tracker [15].”
Diversity Regularized Spatiotemporal Attention for Video-Based Person Re-identification (2018, methodology): “The MARS dataset is the largest video-based person re-identiﬁcation benchmark with 1,261 identities and around 20,000 video sequences generated by DPM detector [10] and GMMCP tracker [7].”
MOTChallenge: A Benchmark for Single-Camera Multiple Target Tracking (2020, methodology): “Conclusion In summary, we observed that after an initial focus on developing algorithms for discrete data association (Dehghan et al. 2015; Le et al. 2016; Pirsiavash et al. 2011; Zhang et al. 2008), the focus shifted towards hand-crafting powerful affinity measures (Choi 2015; Kieritz et al.”