references.bib

@article{computo,
  title = {Computo: reproducible computational/algorithmic contributions in statistics and machine learning},
  author = {{Computo Team}},
  year = {2021},
  journal = {computo},
  url = {https://computo.sfds.asso.fr/}
}

@article{perez2011python,
    title	= {Python: an ecosystem for scientific computing},
    author	= {Perez, Fernando and Granger, Brian E and Hunter, John D},
    journal	= {Computing in Science \\& Engineering},
    volume	= {13},
    number	= {2},
    pages	= {13--21},
    year	= {2011},
    publisher	= {AIP Publishing}
}


@article{garcia2020,
	title = {A metric on the space of finite sets of trajectories for evaluation of multi-target tracking algorithms},
	volume = {68},
	issn = {1053-587X, 1941-0476},
	doi = {10.1109/TSP.2020.3005309},
	abstract = {In this paper, we propose a metric on the space of finite sets of trajectories for assessing multi-target tracking algorithms in a mathematically sound way. The main use of the metric is to compare estimates of trajectories from different algorithms with the ground truth of trajectories. The proposed metric includes intuitive costs associated to localization error for properly detected targets, missed and false targets and track switches at each time step. The metric computation is based on solving a multi-dimensional assignment problem. We also propose a lower bound for the metric, which is also a metric for sets of trajectories and is computable in polynomial time using linear programming. We also extend the proposed metrics on sets of trajectories to random finite sets of trajectories.},
	urldate = {2022-12-13},
	journal = {IEEE Transactions on Signal Processing},
	author = {García-Fernández, Ángel F. and Rahmathullah, Abu Sajana and Svensson, Lennart},
	year = {2020},
	keywords = {Computer Science - Computer Vision and Pattern Recognition, Electrical Engineering and Systems Science - Systems and Control},
	pages = {3917--3928}
}

@ARTICLE{mahler2003,
  author={Mahler, R.P.S.},
  journal={IEEE Transactions on Aerospace and Electronic Systems}, 
  title={Multitarget Bayes filtering via first-order multitarget moments}, 
  year={2003},
  volume={39},
  number={4},
  pages={1152-1178},
  doi={10.1109/TAES.2003.1261119}}

@INPROCEEDINGS{fisher2017,
  author={Yu, Fisher and Wang, Dequan and Shelhamer, Evan and Darrell, Trevor},
  booktitle={2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition}, 
  title={Deep Layer Aggregation}, 
  year={2018},
  volume={},
  number={},
  pages={2403-2412},
  doi={10.1109/CVPR.2018.00255}}

@book{paragios2006,
  title={Handbook of mathematical models in computer vision},
  author={Paragios, Nikos and Chen, Yunmei and Faugeras, Olivier D},
  year={2006},
  publisher={Springer Science \& Business Media}
}

@article{rochman2016,
author = {Rochman, Chelsea and Andrady, Anthony and Dudas, Sarah and Fabres, Joan and Galgani, François and lead, Denise and Hidalgo-Ruz, Valeria and Hong, Sunny and Kershaw, Peter and Lebreton, Laurent and Lusher, Amy and Narayan, Ramani and Pahl, Sabine and Potemra, James and Rochman, Chelsea and Sherif, Sheck and Seager, Joni and Shim, Won and Sobral, Paula and Amaral-Zettler, Linda},
year = {2016},
month = {12},
pages = {},
title = {Sources, fate and effects of microplastics in the marine environment: Part 2 of a global assessment},
journal = {}
}


@article{gonzales2021,
  title={Floating macrolitter leaked from Europe into the ocean},
  author={D. Gonz{\'a}lez-Fern{\'a}ndez and A. C{\'o}zar and G. Hanke and J. Viejo and C. Morales-Caselles and R. Bakiu and D. Barcelo and F. Bessa and Antoine Bruge and M. Cabrera and J. Castro-Jim{\'e}nez and M. Constant and R. Crosti and Yuri Galletti and A. Kideyş and N. Machitadze and Joana Pereira de Brito and M. Pogojeva and N. Ratola and J. Rigueira and E. Rojo-Nieto and O. Savenko and R. I. Sch{\"o}neich-Argent and G. Siedlewicz and Giuseppe Suaria and Myrto Tourgeli},
  journal={Nature Sustainability},
  year={2021},
  volume={4},
  pages={474 - 483}
}


@article{jambeck2015,
author = {Jambeck, Jenna and Geyer, Roland and Wilcox, Chris and Siegler, Theodore and Perryman, Miriam and Andrady, Anthony and Narayan, Ramani and Law, Kara},
year = {2015},
month = {02},
pages = {768-771},
title = {Marine pollution. Plastic waste inputs from land into the ocean},
volume = {347},
journal = {Science (New York, N.Y.)},
doi = {10.1126/science.1260352}
}

@article{welden2020,
author = {Welden, Natalie},
year = {2020},
month = {01},
pages = {195-222},
title = {The environmental impacts of plastic pollution},
isbn = {9780128178805},
doi = {10.1016/B978-0-12-817880-5.00008-6},
journal = {}
}

@article{gamage2020,
author = {Gamage, Thushari and Senevirathna, J.D.M.},
year = {2020},
month = {08},
pages = {e04709},
title = {Plastic pollution in the marine environment},
volume = {6},
journal = {Heliyon},
doi = {10.1016/j.heliyon.2020.e04709}
}

@article{geyer2017,
author = {Geyer, Roland and Jambeck, Jenna and Law, Kara},
year = {2017},
month = {07},
pages = {e1700782},
title = {Production, use, and fate of all plastics ever made},
volume = {3},
journal = {Science Advances},
doi = {10.1126/sciadv.1700782}
}

@book{sarkka2013bayesian,
 author = {S\"arkk\"a, S.},
 title = {Bayesian Filtering and Smoothing},
 year = {2013},
 publisher = {Cambridge University Press},
 address = {New York, NY, USA}
} 

@article{dempster1977maximum,
  title={Maximum likelihood from incomplete data via the EM algorithm},
  author={Dempster, Arthur P and Laird, Nan M and Rubin, Donald B},
  journal={Journal of the Royal Statistical Society: Series B (Methodological)},
  volume={39},
  number={1},
  pages={1--22},
  year={1977},
  publisher={Wiley Online Library}
}

@book{douc2014nonlinear,
  title={Nonlinear time series: Theory, methods and applications with R examples},
  author={Douc, R. and Moulines, \'E. and Stoffer, D.},
  year={2014},
  publisher={CRC press}
}

@inproceedings{farneback2003two,
  title={Two-frame motion estimation based on polynomial expansion},
  author={Farneb{\"a}ck, G.},
  booktitle={Scandinavian conference on Image analysis},
  pages={363--370},
  year={2003},
  organization={Springer}
}


@article{kuhn,
author = {Kuhn, H. W.},
title = {The Hungarian method for the assignment problem},
journal = {Naval Research Logistics Quarterly},
volume = {2},
number = {1-2},
pages = {83-97},
doi = {https://doi.org/10.1002/nav.3800020109},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/nav.3800020109},
eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1002/nav.3800020109},
abstract = {Abstract Assuming that numerical scores are available for the performance of each of n persons on each of n jobs, the “assignment problem” is the quest for an assignment of persons to jobs so that the sum of the n scores so obtained is as large as possible. It is shown that ideas latent in the work of two Hungarian mathematicians may be exploited to yield a new method of solving this problem.},
year = {1955}
}


@proceedings{DBLP:conf/acml/2009,
  editor    = {Zhi{-}Hua Zhou and
               Takashi Washio},
  title     = {Advances in Machine Learning, First Asian Conference on Machine Learning,
               {ACML} 2009, Nanjing, China, November 2-4, 2009. Proceedings},
  series    = {Lecture Notes in Computer Science},
  volume    = {5828},
  publisher = {Springer},
  year      = {2009},
  url       = {http://dx.doi.org/10.1007/978-3-642-05224-8},
  doi       = {10.1007/978-3-642-05224-8},
  isbn      = {978-3-642-05223-1},
  timestamp = {Wed, 11 Nov 2009 10:49:45 +0100},
  biburl    = {http://dblp.uni-trier.de/rec/bib/conf/acml/2009},
  bibsource = {dblp computer science bibliography, http://dblp.org}
}

@proceedings{DBLP:conf/acml/2010,
  editor    = {Masashi Sugiyama and
               Qiang Yang},
  title     = {Proceedings of the 2nd Asian Conference on Machine Learning, {ACML}
               2010, Tokyo, Japan, November 8-10, 2010},
  series    = {{JMLR} Proceedings},
  volume    = {13},
  publisher = {JMLR.org},
  year      = {2010},
  url       = {http://jmlr.org/proceedings/papers/v13/},
  timestamp = {Thu, 11 Sep 2014 07:28:55 +0200},
  biburl    = {http://dblp.uni-trier.de/rec/bib/conf/acml/2010},
  bibsource = {dblp computer science bibliography, http://dblp.org}
}

@proceedings{DBLP:conf/acml/2011,
  editor    = {Chun{-}Nan Hsu and
               Wee Sun Lee},
  title     = {Proceedings of the 3rd Asian Conference on Machine Learning, {ACML}
               2011, Taoyuan, Taiwan, November 13-15, 2011},
  series    = {{JMLR} Proceedings},
  volume    = {20},
  publisher = {JMLR.org},
  year      = {2011},
  url       = {http://jmlr.org/proceedings/papers/v20/},
  timestamp = {Thu, 11 Sep 2014 07:28:55 +0200},
  biburl    = {http://dblp.uni-trier.de/rec/bib/conf/acml/2011},
  bibsource = {dblp computer science bibliography, http://dblp.org}
}

@proceedings{DBLP:conf/acml/2012,
  editor    = {Steven C. H. Hoi and
               Wray L. Buntine},
  title     = {Proceedings of the 4th Asian Conference on Machine Learning, {ACML}
               2012, Singapore, Singapore, November 4-6, 2012},
  series    = {{JMLR} Proceedings},
  volume    = {25},
  publisher = {JMLR.org},
  year      = {2012},
  url       = {http://jmlr.org/proceedings/papers/v25/},
  timestamp = {Thu, 11 Sep 2014 07:28:55 +0200},
  biburl    = {http://dblp.uni-trier.de/rec/bib/conf/acml/2012},
  bibsource = {dblp computer science bibliography, http://dblp.org}
}

@proceedings{DBLP:conf/acml/2013,
  editor    = {Cheng Soon Ong and
               Tu Bao Ho},
  title     = {Asian Conference on Machine Learning, {ACML} 2013, Canberra, ACT,
               Australia, November 13-15, 2013},
  series    = {{JMLR} Proceedings},
  volume    = {29},
  publisher = {JMLR.org},
  year      = {2013},
  url       = {http://jmlr.org/proceedings/papers/v29/},
  timestamp = {Thu, 11 Sep 2014 07:28:55 +0200},
  biburl    = {http://dblp.uni-trier.de/rec/bib/conf/acml/2013},
  bibsource = {dblp computer science bibliography, http://dblp.org}
}



@inproceedings{Liu2018,
abstract = {We present a novel fruit counting pipeline that combines deep segmentation, frame to frame tracking, and 3D localization to accurately count visible fruits across a sequence of images. Our pipeline works on image streams from a monocular camera, both in natural light, as well as with controlled illumination at night. We first train a Fully Convolutional Network (FCN) and segment video frame images into fruit and non-fruit pixels. We then track fruits across frames using the Hungarian Algorithm where the objective cost is determined from a Kalman Filter corrected Kanade-Lucas-Tomasi (KLT) Tracker. In order to correct the estimated count from tracking process, we combine tracking results with a Structure from Motion (SfM) algorithm to calculate relative 3D locations and size estimates to reject outliers and double counted fruit tracks. We evaluate our algorithm by comparing with ground-truth human-annotated visual counts. Our results demonstrate that our pipeline is able to accurately and reliably count fruits across image sequences, and the correction step can significantly improve the counting accuracy and robustness. Although discussed in the context of fruit counting, our work can extend to detection, tracking, and counting of a variety of other stationary features of interest such as leaf-spots, wilt, and blossom.},
archivePrefix = {arXiv},
arxivId = {1804.00307},
author = {Liu, Xu and Chen, Steven W. and Aditya, Shreyas and Sivakumar, Nivedha and Dcunha, Sandeep and Qu, Chao and Taylor, Camillo J. and Das, Jnaneshwar and Kumar, Vijay},
booktitle = {IEEE International Conference on Intelligent Robots and Systems},
doi = {10.1109/IROS.2018.8594239},
eprint = {1804.00307},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Liu et al. - 2018 - Robust Fruit Counting Combining Deep Learning, Tracking, and Structure from Motion.pdf:pdf},
isbn = {9781538680940},
issn = {21530866},
mendeley-groups = {PhD/counting_algorithm/counting_in_videos/with_tracking,PhD/counting_algorithm/counting_in_videos},
pages = {1045--1052},
title = {{Robust Fruit Counting: Combining Deep Learning, Tracking, and Structure from Motion}},
year = {2018}
}


@misc{Zhou2019,
abstract = {Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and requires additional post-processing. In this paper, we take a different approach. We model an object as a single point — the center point of its bounding box. Our detector uses keypoint estimation to find center points and regresses to all other object properties, such as size, 3D location, orientation, and even pose. Our center point based approach, CenterNet, is end-to-end differentiable, simpler, faster, and more accurate than corresponding bounding box based detectors. CenterNet achieves the best speed-accuracy trade-off on the MS COCO dataset, with 28.1% AP at 142 FPS, 37.4% AP at 52 FPS, and 45.1% AP with multi-scale testing at 1.4 FPS. We use the same approach to estimate 3D bounding box in the KITTI benchmark and human pose on the COCO keypoint dataset. Our method performs competitively with sophisticated multi-stage methods and runs in real-time.},
archivePrefix = {arXiv},
arxivId = {1904.07850},
author = {Zhou, Xingyi and Wang, Dequan and Kr{\"{a}}henb{\"{u}}hl, Philipp},
booktitle = {arXiv},
eprint = {1904.07850},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Zhou, Wang, Kr{\"{a}}henb{\"{u}}hl - 2019 - Objects as points.pdf:pdf},
mendeley-groups = {PhD/general_computer_vision/object_detection},
month = {apr},
publisher = {arXiv},
title = {{Objects as points}},
url = {http://arxiv.org/abs/1904.07850},
year = {2019}
}

@misc{Duan,
abstract = {In object detection, keypoint-based approaches often suffer a large number of incorrect object bounding boxes, arguably due to the lack of an additional look into the cropped regions. This paper presents an efficient solution which explores the visual patterns within each cropped region with minimal costs. We build our framework upon a representative one-stage keypoint-based detector named CornerNet. Our approach, named CenterNet, detects each object as a triplet, rather than a pair, of keypoints, which improves both precision and recall. Accordingly, we design two customized modules named cascade corner pooling and center pooling, which play the roles of enriching information collected by both top-left and bottom-right corners and providing more recognizable information at the central regions, respectively. On the MS-COCO dataset, CenterNet achieves an AP of 47.0%, which outperforms all existing one-stage detectors by at least 4.9%. Meanwhile, with a faster inference speed, CenterNet demonstrates quite comparable performance to the top-ranked two-stage detectors.},
author = {Duan, Kaiwen and Bai, Song and Xie, Lingxi and Qi, Honggang and Huang, Qingming and Tian, Qi},
booktitle = {arXiv},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Duan et al. - 2019 - CenterNet Keypoint triplets for object detection.pdf:pdf},
mendeley-groups = {PhD/general_computer_vision/object_detection,PhD/general_computer_vision},
title = {{CenterNet: Keypoint triplets for object detection}},
url = {https://github.com/},
year = {2019}
}


@article{Ciaparrone2020b,
abstract = {The problem of Multiple Object Tracking (MOT) consists in following the trajectory of different objects in a sequence, usually a video. In recent years, with the rise of Deep Learning, the algorithms that provide a solution to this problem have benefited from the representational power of deep models. This paper provides a comprehensive survey on works that employ Deep Learning models to solve the task of MOT on single-camera videos. Four main steps in MOT algorithms are identified, and an in-depth review of how Deep Learning was employed in each one of these stages is presented. A complete experimental comparison of the presented works on the three MOTChallenge datasets is also provided, identifying a number of similarities among the top-performing methods and presenting some possible future research directions.},
archivePrefix = {arXiv},
arxivId = {1907.12740},
author = {Ciaparrone, Gioele and {Luque S{\'{a}}nchez}, Francisco and Tabik, Siham and Troiano, Luigi and Tagliaferri, Roberto and Herrera, Francisco},
doi = {10.1016/j.neucom.2019.11.023},
eprint = {1907.12740},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Ciaparrone et al. - 2020 - Deep learning in video multi-object tracking A survey.pdf:pdf},
issn = {18728286},
journal = {Neurocomputing},
keywords = {Convolutional neural networks,Deep learning,LSTM,Multiple object tracking,Reinforcement learning,Video tracking},
mendeley-groups = {PhD/multi_object_tracking,PhD/multi_object_tracking/reviews},
pages = {61--88},
title = {{Deep learning in video multi-object tracking: A survey}},
volume = {381},
year = {2020}
}

@article{Zhanga,
  title={Fairmot: On the fairness of detection and re-identification in multiple object tracking},
  author={Zhang, Yifu and Wang, Chunyu and Wang, Xinggang and Zeng, Wenjun and Liu, Wenyu},
  journal={International Journal of Computer Vision},
  pages={1--19},
  year={2021},
  publisher={Springer}
}

@article{zhou2020,
author = {Zhou, Xingyi and Koltun, Vladlen and Krähenbühl, Philipp},
year = {2020},
month = {10},
pages = {474-490},
title = {Tracking Objects as Points},
isbn = {978-3-030-58547-1},
doi = {10.1007/978-3-030-58548-8_28},
journal = {}
}
@article{Lebreton2017,
abstract = {Plastics in the marine environment have become a major concern because of their persistence at sea, and adverse consequences to marine life and potentially human health. Implementing mitigation strategies requires an understanding and quantification of marine plastic sources, taking spatial and temporal variability into account. Here we present a global model of plastic inputs from rivers into oceans based on waste management, population density and hydrological information. Our model is calibrated against measurements available in the literature. We estimate that between 1.15 and 2.41 million tonnes of plastic waste currently enters the ocean every year from rivers, with over 74% of emissions occurring between May and October. The top 20 polluting rivers, mostly located in Asia, account for 67% of the global total. The findings of this study provide baseline data for ocean plastic mass balance exercises, and assist in prioritizing future plastic debris monitoring and mitigation strategies.},
author = {Lebreton, Laurent C.M. and {Van Der Zwet}, Joost and Damsteeg, Jan Willem and Slat, Boyan and Andrady, Anthony and Reisser, Julia},
doi = {10.1038/ncomms15611},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Lebreton et al. - 2017 - River plastic emissions to the world's oceans.pdf:pdf},
issn = {20411723},
journal = {Nature Communications},
mendeley-groups = {PhD/plastic_pollution},
pages = {1--10},
pmid = {28589961},
publisher = {Nature Publishing Group},
title = {{River plastic emissions to the world's oceans}},
url = {http://dx.doi.org/10.1038/ncomms15611},
volume = {8},
year = {2017}
}
@article{Bruge2018,
abstract = {Rivers are major pathways for litter to enter the ocean, especially plastic debris. Yet, further research is needed to improve knowledge on rivers contribution, increase data availability, refine litter origins, and develop relevant solutions to limit riverine litter inputs. This study presents the results of three years of aquatic litter monitoring on the Adour river catchment (southwest of France). Litter monitoring consisted of collecting all litter stranded on river banks or stuck in the riparian vegetation in defined areas identified from cartographic and hydromorphological analyses, and with the support of local stakeholders. Litter samples were then sorted and counted according to a list of items containing 130 categories. Since 2014, 278 litter samplings were carried out, and 120,632 litter items were collected, sorted, and counted. 41% of litter could not be identified due to high degradation. Food and beverage packaging, smoking-related items, sewage related debris, fishery and mariculture gear, and common household items represented around 70% of identifiable items. Overall, the present study contributes to our knowledge of litter sources and pathways, with the target of reducing the amounts entering the ocean. The long-term application of this monitoring is a way forward to measure societal changes as well as assess effectiveness of measures.},
author = {Bruge, Antoine and Barreau, Cristina and Carlot, J{\'{e}}r{\'{e}}my and Collin, H{\'{e}}l{\`{e}}ne and Moreno, Cl{\'{e}}ment and Maison, Philippe},
doi = {10.3390/jmse6010024},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Bruge et al. - 2018 - Monitoring litter inputs from the Adour river (southwest France) to the marine environment.pdf:pdf},
issn = {20771312},
journal = {Journal of Marine Science and Engineering},
keywords = {Aquatic litter,Monitoring,Plastic pollution,Pollution,Riverine input},
mendeley-groups = {PhD/plastic_pollution},
number = {1},
title = {{Monitoring litter inputs from the Adour river (southwest France) to the marine environment}},
volume = {6},
year = {2018}
}
@article{VanEmmerik2019,
abstract = {Rivers transport land-based plastic waste into the ocean. Current efforts to quantify riverine plastic emission come with uncertainty as field observations are scarce. One of the challenging aspects is the lack of consistent measurement methods that allow for comparing rivers over space and time. Recent studies have shown that simple visual observations provide a robust first-order characterization of floating and superficially suspended plastic transport, both in quantity, spatiotemporal distribution and composition. For this study, we applied this method to the river Seine, France, to provide new insights in the spatiotemporal variation in riverine plastic transport. First, we studied the response of plastic flow to increased river discharge by comparing measurements taken during low flow and high flow periods. Second, we investigated the variation of riverine plastic transport over the river length to improve our understanding of the origin and fate of riverine plastics. We demonstrate that during a period with higher river discharge, plastic transport increased up to a factor ten at the observation point closest to the river mouth. This suggests that the plastic emission into the ocean from the Seine may also be considerably higher during increased discharge. Upstream of Paris plastic transport increased only with a factor 1.5, suggesting that most plastics originate from Paris or areas further downstream. With this paper we aim to shed additional light on the seasonal variation in riverine plastic transport and its distribution along the river length, which may benefit future long-term monitoring efforts and plastic pollution mitigation strategies.},
author = {van Emmerik, Tim and Tramoy, Romain and van Calcar, Caroline and Alligant, Soline and Treilles, Robin and Tassin, Bruno and Gasperi, Johnny},
doi = {10.3389/fmars.2019.00642},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/van Emmerik et al. - 2019 - Seine Plastic Debris Transport Tenfolded During Increased River Discharge.pdf:pdf},
issn = {22967745},
journal = {Frontiers in Marine Science},
keywords = {France,Seine,hydrology,marine plastic debris,plastic pollution,plastic pollution monitoring,river plastic},
mendeley-groups = {PhD/plastic_pollution},
number = {October},
pages = {1--7},
title = {{Seine Plastic Debris Transport Tenfolded During Increased River Discharge}},
volume = {6},
year = {2019}
}

@article{VanEmmerik2020,
author = {van Emmerik, Tim and Schwarz, Anna},
title = {Plastic debris in rivers},
journal = {WIREs Water      },
volume = {7},
number = {1},
pages = {e1398},
keywords = {hydrology, macroplastic, marine plastic litter, microplastic, riverine plastic pollution},
doi = {https://doi.org/10.1002/wat2.1398},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/wat2.1398},
eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1002/wat2.1398},
abstract = {Abstract Plastic pollution in aquatic ecosystems is an emerging environmental risk, as it may negatively impacts ecology, endangers aquatic species, and causes economic damage. Rivers are known to play a crucial role in transporting land-based plastic waste to the world's oceans, but riverine ecosystems are also directly affected by plastic pollution. To better quantify global plastic pollution transport and to effectively reduce sources and risks, a thorough understanding of origin, transport, fate, and effects of riverine plastic debris is crucial. In this overview paper, we discuss the current scientific state on plastic debris in rivers and evaluate existing knowledge gaps. We present a brief background of plastics, polymer types typically found in rivers, and the risk posed to aquatic ecosystems. Additionally, we elaborate on the origin and fate of riverine plastics, including processes and factors influencing plastic debris transport and its spatiotemporal variation. We present an overview of monitoring and modeling efforts to characterize riverine plastic transport, and give examples of typical values from around the world. Finally, we present an outlook to riverine plastic research. With this paper, we aim to present an inclusive and comprehensive overview of riverine plastic debris research to date and suggest multiple ways forward for future research. This article is categorized under: Science of Water > Water Quality Water and Life > Stresses and Pressures on Ecosystems},
year = {2020}
}


@article{VanEmmerik2019b,
  
AUTHOR={van Emmerik, Tim and Tramoy, Romain and van Calcar, Caroline and Alligant, Soline and Treilles, Robin and Tassin, Bruno and Gasperi, Johnny},   
	 
TITLE={Seine Plastic Debris Transport Tenfolded During Increased River Discharge},      
	
JOURNAL={Frontiers in Marine Science},      
	
VOLUME={6},      

PAGES={642},     
	
YEAR={2019},      
	  
URL={https://www.frontiersin.org/article/10.3389/fmars.2019.00642},       
	
DOI={10.3389/fmars.2019.00642},      
	
ISSN={2296-7745},   
   
ABSTRACT={Rivers transport land-based plastic waste into the ocean. Current efforts to quantify riverine plastic emission come with uncertainty as field observations are scarce. One of the challenging aspects is the lack of consistent measurement methods that allow for comparing rivers over space and time. Recent studies have shown that simple visual observations provide a robust first-order characterization of floating and superficially suspended plastic transport, both in quantity, spatiotemporal distribution and composition. For this study, we applied this method to the river Seine, France, to provide new insights in the spatiotemporal variation in riverine plastic transport. First, we studied the response of plastic flow to increased river discharge by comparing measurements taken during low flow and high flow periods. Second, we investigated the variation of riverine plastic transport over the river length to improve our understanding of the origin and fate of riverine plastics. We demonstrate that during a period with higher river discharge, plastic transport increased up to a factor ten at the observation point closest to the river mouth. This suggests that the plastic emission into the ocean from the Seine may also be considerably higher during increased discharge. Upstream of Paris plastic transport increased only with a factor 1.5, suggesting that most plastics originate from Paris or areas further downstream. With this paper we aim to shed additional light on the seasonal variation in riverine plastic transport and its distribution along the river length, which may benefit future long-term monitoring efforts and plastic pollution mitigation strategies.}
}

@article{Castro-Jimenez2019,
abstract = {We present here the first estimates of floating macro-litter in surface waters from the Rhone River, based on monthly visual observations during 1-year period (2016–2017). Plastic represented 77% of the identified items, confirming its predominance in riverine floating litter. Fragments (2.5–50 cm) and Single Use Plastics (i.e. bags, bottles and cover/packaging) were among the most abundant items. Frequent non-plastic floating litter were paper items such as packaging material and newspapers, and metal items (mostly cans), representing 14% and 5% of total litter, respectively. A lower-end estimate resulted in ∼223,000 plastic items (∼0.7 t of plastic) transported annually by the Rhone surface waters to the Gulf of Lion (NW Mediterranean Sea). Floating macro-plastics are only a fraction of the total plastic export by the Rhone. Our study highlights the current discrepancy between field observations and theoretical estimations. Improvements are needed to harmonize data collection methodologies for field studies and model validation.},
author = {Castro-Jim{\'{e}}nez, Javier and Gonz{\'{a}}lez-Fern{\'{a}}ndez, Daniel and Fornier, Michel and Schmidt, Natascha and Semp{\'{e}}r{\'{e}}, Richard},
doi = {10.1016/j.marpolbul.2019.05.067},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Castro-Jim{\'{e}}nez et al. - 2019 - Macro-litter in surface waters from the Rhone River Plastic pollution and loading to the NW Mediterranea.pdf:pdf},
issn = {18793363},
journal = {Marine Pollution Bulletin},
keywords = {Large European rivers,Litter flow,Marine pollution,Plastic waste,Water quality},
mendeley-groups = {PhD/plastic_pollution},
number = {May},
pages = {60--66},
pmid = {31426199},
publisher = {Elsevier},
title = {{Macro-litter in surface waters from the Rhone River: Plastic pollution and loading to the NW Mediterranean Sea}},
url = {https://doi.org/10.1016/j.marpolbul.2019.05.067},
volume = {146},
year = {2019}
}
@article{Lau2012,
abstract = {Video technology has been playing an increasing role in marine science, both for habitat mapping and estimating commercial species abundance. However, when quantification is needed, it is usually based on manual counting, a subjective and time-consuming task. The present work proposes a methodology to automatically quantify the abundance of Norway lobsters, Nephrops norvegicus, by counting lobsters or their burrows from video sequences, as a reliable complement to the currently used operator-based approach. The methodology is validated using a set of test video sequences captured at the Portuguese continental slope, using a monochrome camera mounted on a trawl gear, being characterised by non-uniform illumination, artefacts at image border, noise and marine snow. The analysis includes, after a pre-processing stage, the segmentation of regions of interest and the corresponding classification into one of the three targeted classes: Norway lobsters, burrows and others (including trawl impact marks). The developed software prototype, named IT-IPIMAR N. norvegicus (I2N2), is able to provide an objective, detailed and comprehensive analysis to complement manual evaluation, for lobster and burrow density estimation. {\textcopyright} 2012 The Institution of Engineering and Technology.},
author = {Lau, P. Y. and Correia, P. L. and Fonseca, P. and Campos, A.},
doi = {10.1049/iet-ipr.2009.0426},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Lau et al. - 2012 - Estimating Norway lobster abundance from deep-water videos An automatic approach.pdf:pdf},
issn = {17519659},
journal = {IET Image Processing},
mendeley-groups = {PhD/plastic_pollution},
number = {1},
pages = {22--30},
title = {{Estimating Norway lobster abundance from deep-water videos: An automatic approach}},
volume = {6},
year = {2012}
}

@article{panwar2020aqua,
abstract = {Water pollution is one of serious threats in the society. More than 8 million tons of plastic are dumped in the oceans each year. In addition to that beaches are littered by tourists and residents all around the world. It is no secret that the aquatic life ecosystem is at a risk and soon the ratio of plastic to fish will be 1:1. In this paper, we have proposed a data set known as AquaTrash which is based on TACO data set. Further, we have applied proposed state-of-the-art deep learning-based object detection model known as AquaVision over AquaTrash dataset. Proposed model detects and classifies the different pollutants and harmful waste items floating in the oceans and on the seashores with mean Average Precision (mAP) of 0.8148. The propose method localizes the waste object that help in cleaning the water bodies and contributes to environment by maintaining the aquatic ecosystem.},
author = {Panwar, Harsh and Gupta, P K and Siddiqui, Mohammad Khubeb and Morales-Menendez, Ruben and Bhardwaj, Prakhar and Sharma, Sudhansh and Sarker, Iqbal H},
doi = {https://doi.org/10.1016/j.cscee.2020.100026},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Panwar et al. - 2020 - AquaVision Automating the detection of waste in water bodies using deep transfer learning.pdf:pdf},
issn = {2666-0164},
journal = {Case Studies in Chemical and Environmental Engineering},
keywords = {AquaTrash,Deep learning,RetinaNet,Water pollution,Water waste detection},
mendeley-groups = {PhD/counting_algorithm/counting_in_images/plastic-related},
pages = {100026},
title = {{AquaVision: Automating the detection of waste in water bodies using deep transfer learning}},
url = {http://www.sciencedirect.com/science/article/pii/S2666016420300244},
year = {2020}
}
@article{Wolf2020,
abstract = {Large quantities of mismanaged plastic waste are polluting and
threatening the health of the blue planet. Vast amounts of this
plastic waste found in the oceans originates from land. It finds
its way to the open ocean through rivers, waterways and
estuarine systems. Here we present a novel machine learning
algorithm based on convolutional neural networks (CNNs) that is
capable of detecting and quantifying floating and washed ashore
plastic litter. The aquatic plastic litter detector and
quantifier system (APLASTIC--Q) was developed and trained using
very high geo--spatial resolution imagery (5 pixels/cm = 0.002
m/pixel) captured from aerial surveys in Cambodia. APLASTIC--Q
comprises two machine learning algorithms components (i) plastic
litter detector (PLD--CNN) and (ii) plastic litter quantifier
(PLQ--CNN). PLD--CNN managed to categorize targets as water,
sand, vegetation and plastic litter with an 83 % accuracy. It
also provided a qualitative count of litter as low or high based
on a thresholding approach. PLQ--CNN further distinguished and
enumerated the litter items in each of the classes define as
water bottles, Styrofoam, canisters, cartons, bowls, shoes,
polystyrene packaging, cups, textile, carry bags small or large.
The types and amounts of plastic litter provide benchmark
information that is urgently needed for decision making by
policymakers, citizens and stakeholders especially for
developing plastic policies. Quasi--quantification was based on
automated counts of items present in the imagery with caveats of
underlying object in case of aggregated litter. Our scientific
evidence--based algorithm based on machine learning complement
net trawl surveys, field campaigns and clean--up activities for
improved quantification of plastic litter. APLASTIC--Q will be
an open--source smart algorithm that is easy to adapt for fast
and automated detection as well as quantification of floating or
washed ashore plastic litter from aerial, high--altitude pseudo
satellites and space missions.},
author = {Wolf, Mattis and van den Berg, Katelijn and Garaba, Shungudzemwoyo Pascal and Gnann, Nina and Sattler, Klaus and Stahl, Frederic Theodor and Zielinski, Oliver},
doi = {10.1088/1748-9326/abbd01},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Wolf et al. - 2020 - Machine learning for aquatic plastic litter detection, classification and quantification (APLASTIC–Q).pdf:pdf},
journal = {Environmental Research Letters},
keywords = {convolutional neural networks,detection,machine learning,plastic litter,remote sensing,river and beach ecosystems},
mendeley-groups = {PhD/counting_algorithm/counting_in_images/plastic-related},
title = {{Machine learning for aquatic plastic litter detection, classification and quantification (APLASTIC–Q)}},
year = {2020}
}

@inproceedings{Chattopadhyay,
abstract = {We are interested in counting the number of instances of object classes in natural, everyday images. Previous counting approaches tackle the problem in restricted domains such as counting pedestrians in surveillance videos. Counts can also be estimated from outputs of other vision tasks like object detection. In this work, we build dedicated models for counting designed to tackle the large variance in counts, appearances, and scales of objects found in natural scenes. Our approach is inspired by the phenomenon of subitizing -the ability of humans to make quick assessments of counts given a perceptual signal, for small count values. Given a natural scene, we employ a divide and conquer strategy while incorporating context across the scene to adapt the subitizing idea to counting. Our approach offers consistent improvements over numerous baseline approaches for counting on the PASCAL VOC 2007 and COCO datasets. Subsequently, we study how counting can be used to improve object detection. We then show a proof of concept application of our counting methods to the task of Visual Question Answering, by studying the 'how many?' questions in the VQA and COCO-QA datasets.},
archivePrefix = {arXiv},
arxivId = {1604.03505},
author = {Chattopadhyay, Prithvijit and Vedantam, Ramakrishna and Selvaraju, Ramprasaath R and Batra, Dhruv and Parikh, Devi},
booktitle = {Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017},
doi = {10.1109/CVPR.2017.471},
eprint = {1604.03505},
file = {:home/mmip/.local/share/data/Mendeley Ltd./Mendeley Desktop/Downloaded/Chattopadhyay et al. - 2017 - Counting everyday objects in everyday scenes.pdf:pdf},
isbn = {9781538604571},
mendeley-groups = {PhD/counting_algorithm/counting_in_images},
pages = {4428--4437},
title = {{Counting everyday objects in everyday scenes}},
volume = {2017-Janua},
year = {2017}
}

@inproceedings{Law,
  author    = {Hei Law and
               Jia Deng},
  editor    = {Vittorio Ferrari and
               Martial Hebert and
               Cristian Sminchisescu and
               Yair Weiss},
  title     = {CornerNet: Detecting Objects as Paired Keypoints},
  booktitle = {Computer Vision - {ECCV} 2018 - 15th European Conference, Munich,
               Germany, September 8-14, 2018, Proceedings, Part {XIV}},
  series    = {Lecture Notes in Computer Science},
  volume    = {11218},
  pages     = {765--781},
  publisher = {Springer},
  year      = {2018},
  url       = {https://doi.org/10.1007/978-3-030-01264-9\_45},
  doi       = {10.1007/978-3-030-01264-9\_45},
  timestamp = {Tue, 02 Feb 2021 12:07:19 +0100},
  biburl    = {https://dblp.org/rec/conf/eccv/LawD18.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}

@article{lechner2014,
title = "The Danube so colourful: A potpourri of plastic litter outnumbers fish larvae in Europe's second largest river",
journal = "Environmental Pollution",
volume = "188",
pages = "177 - 181",
year = "2014",
issn = "0269-7491",
doi = "https://doi.org/10.1016/j.envpol.2014.02.006",
url = "http://www.sciencedirect.com/science/article/pii/S0269749114000475",
author = "Aaron Lechner and Hubert Keckeis and Franz Lumesberger-Loisl and Bernhard Zens and Reinhard Krusch and Michael Tritthart and Martin Glas and Elisabeth Schludermann",
keywords = "Plastic debris, Freshwater pollution, Black Sea, Industrial plastics, Drift",
abstract = "Previous studies on plastic pollution of aquatic ecosystems focused on the world's oceans. Large rivers as major pathways for land-based plastic litter, has received less attention so far. Here we report on plastic quantities in the Austrian Danube. A two year survey (2010, 2012) using stationary driftnets detected mean plastic abundance (n = 17,349; mean ± S.D: 316.8 ± 4664.6 items per 1000 m−3) and mass (4.8 ± 24.2 g per 1000 m−3) in the river to be higher than those of drifting larval fish (n = 24,049; 275.3 ± 745.0 individuals. 1000 m−3 and 3.2 ± 8.6 g 1000 m−3). Industrial raw material (pellets, flakes and spherules) accounted for substantial parts (79.4%) of the plastic debris. The plastic input via the Danube into the Black Sea was estimated to 4.2 t per day."
}


@article{gasperi2014,
title = "Assessment of floating plastic debris in surface water along the Seine River",
journal = "Environmental Pollution",
volume = "195",
pages = "163 - 166",
year = "2014",
issn = "0269-7491",
doi = "https://doi.org/10.1016/j.envpol.2014.09.001",
url = "http://www.sciencedirect.com/science/article/pii/S0269749114003807",
author = "Johnny Gasperi and Rachid Dris and Tiffany Bonin and Vincent Rocher and Bruno Tassin",
keywords = "Floating plastic litter, Plastic loads, River, Riverine litter",
abstract = "This study is intended to examine the quality and quantity of floating plastic debris in the River Seine through use of an extensive regional network of floating debris-retention booms; it is one of the first attempts to provide reliable information on such debris at a large regional scale. Plastic debris represented between 0.8\% and 5.1\% of total debris collected by weight. A significant proportion consisted of food wrappers/containers and plastic cutlery, probably originating from voluntary or involuntary dumping, urban discharges and surface runoff. Most plastic items are made of polypropylene, polyethylene and, to a lesser extent, polyethylene terephthalate. By extrapolation, some 27 tons of floating plastic debris are intercepted annually by this network; corresponding to 2.3 g per Parisian inhabitant per year. Such data could serve to provide a first evaluation of floating plastic inputs conveyed by rivers."
}


@article{moritt2014,
title = {Plastic in the Thames: A river runs through it},
journal = {Marine Pollution Bulletin},
volume = {78},
number = {1},
pages = {196-200},
year = {2014},
issn = {0025-326X},
doi = {https://doi.org/10.1016/j.marpolbul.2013.10.035},
url = {https://www.sciencedirect.com/science/article/pii/S0025326X13006565},
author = {David Morritt and Paris V. Stefanoudis and Dave Pearce and Oliver A. Crimmen and Paul F. Clark},
keywords = {Estuary, Fyke-nets, Plastics, River Thames, Sanitary products, United Kingdom},
abstract = {Although contamination of the marine ecosystems by plastics is becoming recognised as a serious pollution problem, there are few studies that demonstrate the contribution made by freshwater catchments. Over a three month period from September to December 2012, at seven localities in the upper Thames estuary, 8490 submerged plastic items were intercepted in eel fyke nets anchored to the river bed. Whilst there were significant differences in the numbers of items at these locations, the majority were some type of plastic. Additionally in excess of 20% of the litter items were components of sanitary products. The most contaminated sites were in the vicinity of sewage treatment works. While floating litter is visible, this study also demonstrates that a large unseen volume of submerged plastic is flowing into the marine environment. It is therefore important that this sub-surface component is considered when assessing plastic pollution input into the sea.}
}

@article{vanlieshout2020automated,
abstract = {Abstract Quantifying plastic pollution on surface water is essential to understand and mitigate the impact of plastic pollution to the environment. Current monitoring methods such as visual counting are labor intensive. This limits the feasibility of scaling to long-term monitoring at multiple locations. We present an automated method for monitoring plastic pollution that overcomes this limitation. Floating macroplastics are detected from images of the water surface using deep learning. We perform an experimental evaluation of our method using images from bridge-mounted cameras at five different river locations across Jakarta, Indonesia. The four main results of the experimental evaluation are as follows. First, we realize a method that obtains a reliable estimate of plastic density (68.7{\%} precision). Our monitoring method successfully distinguishes plastics from environmental elements, such as water surface reflection and organic waste. Second, when trained on one location, the method generalizes well to new locations with relatively similar conditions without retraining (≈50{\%} average precision). Third, generalization to new locations with considerably different conditions can be boosted by retraining on only 50 objects of the new location (improving precision from ≈20{\%} to ≈42{\%}). Fourth, our method matches visual counting methods and detects ≈35{\%} more plastics, even more so during periods of plastic transport rates of above 10 items per meter per minute. Taken together, these results demonstrate that our method is a promising way of monitoring plastic pollution. By extending the variety of the data set the monitoring method can be readily applied at a larger scale.},
annote = {From Duplicate 1 (Automated River Plastic Monitoring Using Deep Learning and Cameras - van Lieshout, Colin; van Oeveren, Kees; van Emmerik, Tim; Postma, Eric)

e2019EA000960 10.1029/2019EA000960

From Duplicate 2 (Automated River Plastic Monitoring Using Deep Learning and Cameras - van Lieshout, Colin; van Oeveren, Kees; van Emmerik, Tim; Postma, Eric)

From Duplicate 2 (Automated River Plastic Monitoring Using Deep Learning and Cameras - van Lieshout, Colin; van Oeveren, Kees; van Emmerik, Tim; Postma, Eric)

e2019EA000960 10.1029/2019EA000960},
author = {van Lieshout, Colin and van Oeveren, Kees and van Emmerik, Tim and Postma, Eric},
doi = {10.1029/2019EA000960},
issn = {23335084},
journal = {Earth and Space Science},
keywords = {artificial intelligence,automated monitoring,deep learning,object detection,plastic pollution,river plastic},
number = {8},
pages = {e2019EA000960},
title = {{Automated River Plastic Monitoring Using Deep Learning and Cameras}},
volume = {7},
year = {2020}
}

@techreport{Gao2003,
abstract = {Accurately estimating the number of objects in a single image is a challenging yet meaningful task and has been applied in many applications such as urban planning and public safety. In the various object counting tasks, crowd counting is particularly prominent due to its specific significance to social security and development. Fortunately, the development of the techniques for crowd counting can be generalized to other related fields such as vehicle counting and environment survey, if without taking their characteristics into account. Therefore, many researchers are devoting to crowd counting, and many excellent works of literature and works have spurted out. In these works, they are must be helpful for the development of crowd counting. However, the question we should consider is why they are effective for this task. Limited by the cost of time and energy, we cannot analyze all the algorithms. In this paper, we have surveyed over 220 works to comprehensively and systematically study the crowd counting models, mainly CNN-based density map estimation methods. Finally, according to the evaluation metrics, we select the top three performers on their crowd counting datasets and analyze their merits and drawbacks. Through our analysis, we expect to make reasonable inference and prediction for the future development of crowd counting, and meanwhile, it can also provide feasible solutions for the problem of object counting in other fields. We provide the density maps and prediction results of some mainstream algorithm in the validation set of NWPU dataset for comparison and testing. Meanwhile, density map generation and evaluation tools are also provided. All the codes and evaluation results are made publicly available at https://github.com/gaoguangshuai/survey-for-crowd-counting.},
archivePrefix = {arXiv},
arxivId = {2003.12783v1},
author = {Gao, Guangshuai and Gao, Junyu and Liu, Qingjie and Wang, Qi and Wang, Yunhong},
eprint = {2003.12783v1},
keywords = {CNNs,Index Terms-Object counting,crowd counting,density estimation},
title = {{CNN-based Density Estimation and Crowd Counting: A Survey}},
url = {https://github.com/gaoguangshuai/},
year = {2003}
}



@article{Arteta2016,
author = {Arteta, Carlos and Lempitsky, Victor and Zisserman, Andrew},
year = {2016},
month = {10},
pages = {483-498},
title = {Counting in the Wild},
volume = {9911},
isbn = {978-3-319-46477-0},
doi = {10.1007/978-3-319-46478-7_30},
journal = {}
}

@article{wu2020fast,
title = {Fast video crowd counting with a Temporal Aware Network},
journal = {Neurocomputing},
volume = {403},
pages = {13-20},
year = {2020},
issn = {0925-2312},
doi = {https://doi.org/10.1016/j.neucom.2020.04.071},
url = {https://www.sciencedirect.com/science/article/pii/S0925231220306561},
author = {Xingjiao Wu and Baohan Xu and Yingbin Zheng and Hao Ye and Jing Yang and Liang He},
keywords = {Crowd counting, Video analysis, Spatiotemporal information, Temporal Aware Network},
abstract = {Crowd counting aims to count the number of instantaneous people in a crowded space, and many promising solutions have been proposed for single image crowd counting. With the ubiquitous video capture devices in public safety field, how to effectively apply the crowd counting technique to video content has become an urgent problem. In this paper, we introduce a novel framework based on temporal aware modeling of the relationship between video frames. The proposed network contains a few dilated residual blocks, and each of them consists of the layers that compute the temporal convolutions of features from the adjacent frames to improve the prediction. To alleviate the expensive computation and satisfy the demand of fast video crowd counting, we also introduce a lightweight network to balance the computational cost with representation ability. We conduct experiments on the crowd counting benchmarks and demonstrate its superiority in terms of effectiveness and efficiency over previous video-based approaches.}
}


@inproceedings{ren2016faster,
author = {Ren, Shaoqing and He, Kaiming and Girshick, Ross and Sun, Jian},
title = {Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks},
year = {2015},
publisher = {MIT Press},
address = {Cambridge, MA, USA},
booktitle = {Proceedings of the 28th International Conference on Neural Information Processing Systems - Volume 1},
pages = {91–99},
numpages = {9},
location = {Montreal, Canada},
series = {NIPS'15}
}


@article{Proenca2020,
abstract = {TACO is an open image dataset for litter detection and segmentation, which is growing through crowdsourcing. Firstly, this paper describes this dataset and the tools developed to support it. Secondly, we report instance segmentation performance using Mask R-CNN on the current version of TACO. Despite its small size (1500 images and 4784 annotations), our results are promising on this challenging problem. However, to achieve satisfactory trash detection in the wild for deployment, TACO still needs much more manual annotations. These can be contributed using: http://tacodataset.org/},
archivePrefix = {arXiv},
arxivId = {2003.06975},
author = {Proen{\c{c}}a, Pedro F and Sim{\~{o}}es, Pedro},
eprint = {2003.06975},
title = {{TACO: Trash Annotations in Context for Litter Detection}},
url = {http://tacodataset.org/ http://arxiv.org/abs/2003.06975},
year = {2020},
journal = {}
}


@INPROCEEDINGS {bergmann2019,
author = {P. Bergmann and T. Meinhardt and L. Leal-Taixe},
booktitle = {2019 IEEE/CVF International Conference on Computer Vision (ICCV)},
title = {Tracking Without Bells and Whistles},
year = {2019},
volume = {},
issn = {},
pages = {941-951},
keywords = {detectors;trajectory;target tracking;object detection;task analysis;proposals},
doi = {10.1109/ICCV.2019.00103},
url = {https://doi.ieeecomputersociety.org/10.1109/ICCV.2019.00103},
publisher = {IEEE Computer Society},
address = {Los Alamitos, CA, USA},
month = {nov}
}



@article{luiten2020,
  title={Hota: A higher order metric for evaluating multi-object tracking},
  author={Luiten, Jonathon and Osep, Aljosa and Dendorfer, Patrick and Torr, Philip and Geiger, Andreas and Leal-Taix{\'e}, Laura and Leibe, Bastian},
  journal={International journal of computer vision},
  volume={129},
  number={2},
  pages={548--578},
  year={2021},
  publisher={Springer}
}


@inproceedings{Fulton2018,
  title={Robotic detection of marine litter using deep visual detection models},
  author={Fulton, Michael and Hong, Jungseok and Islam, Md Jahidul and Sattar, Junaed},
  booktitle={2019 International Conference on Robotics and Automation (ICRA)},
  pages={5752--5758},
  year={2019},
  organization={IEEE}
}

@inproceedings{Hong2020,
abstract = {This paper presents an approach to address data scarcity problems in underwater image datasets for visual detection of marine debris. The proposed approach relies on a two-stage variational autoencoder (VAE) and a binary classifier to evaluate the generated imagery for quality and realism. From the images generated by the two-stage VAE, the binary classifier selects good quality images and augments the given dataset with them. Lastly, a multi-class classifier is used to evaluate the impact of the augmentation process by measuring the accuracy of an object detector trained on combinations of real and generated trash images. Our results show that the classifier trained with the augmented data outperforms the one trained only with the real data. This approach will not only be valid for the underwater trash classification problem presented in this paper, but it will also be useful for any data-dependent task for which collecting more images is challenging or infeasible.},
archivePrefix = {arXiv},
arxivId = {1910.04754},
author = {Hong, Jungseok and Fulton, Michael and Sattar, Junaed},
booktitle = {Proceedings - IEEE International Conference on Robotics and Automation},
doi = {10.1109/ICRA40945.2020.9197575},
eprint = {1910.04754},
isbn = {9781728173955},
issn = {10504729},
pages = {10525--10531},
title = {{A Generative Approach Towards Improved Robotic Detection of Marine Litter}},
year = {2020}
}


@INPROCEEDINGS{Xiong2017,
  author={Xiong, Feng and Shi, Xingjian and Yeung, Dit-Yan},
  booktitle={2017 IEEE International Conference on Computer Vision (ICCV)}, 
  title={Spatiotemporal Modeling for Crowd Counting in Videos}, 
  year={2017},
  volume={},
  number={},
  pages={5161-5169},
  doi={10.1109/ICCV.2017.551}}


@article{Miao2019,
abstract = {The task of crowd counting and density maps estimating from videos is challenging due to severe occlusions, scene perspective distortions and diverse crowd distributions. Conventional crowd counting methods via deep learning technique process each video frame independently with no consideration of the intrinsic temporal correlation among neighboring frames, thus making the performance lower than the required level of real-world applications. To overcome this shortcoming, a new end-to-end deep architecture named Spatial-Temporal Convolutional Neural Network (ST-CNN) is proposed, which unifies 2D convolutional neural network (C2D) and 3D convolutional neural network (C3D) to learn spatial-temporal features in the same framework. On top of that, a merging scheme is performed on the resulting density maps, taking advantages of the spatial-temporal information simultaneously for the crowd counting task. Experimental results on two benchmark data sets {\^{a}} Mall dataset and WorldExpo′10 dataset show that our ST-CNN outperforms the state-of-the-art models in terms of mean absolutely error (MAE) and mean squared error (MSE).},
author = {Miao, Yunqi and Han, Jungong and Gao, Yongsheng and Zhang, Baochang},
doi = {10.1016/j.patrec.2019.04.012},
issn = {01678655},
journal = {Pattern Recognition Letters},
keywords = {Crowd analysis,Crowd counting,Spatio-temporal feature},
mendeley-groups = {PhD/counting_algorithm/counting_in_videos/with_rnn_or_conv3D,PhD/counting_algorithm/counting_in_videos},
month = {jul},
pages = {113--118},
publisher = {Elsevier B.V.},
title = {{ST-CNN: Spatial-Temporal Convolutional Neural Network for crowd counting in videos}},
volume = {125},
year = {2019}
}

@article{luo2021,
title = {Multiple object tracking: A literature review},
journal = {Artificial Intelligence},
volume = {293},
pages = {103448},
year = {2021},
issn = {0004-3702},
doi = {https://doi.org/10.1016/j.artint.2020.103448},
url = {https://www.sciencedirect.com/science/article/pii/S0004370220301958},
author = {Wenhan Luo and Junliang Xing and Anton Milan and Xiaoqin Zhang and Wei Liu and Tae-Kyun Kim},
keywords = {Multi-object tracking, Data association, Survey},
abstract = {Multiple Object Tracking (MOT) has gained increasing attention due to its academic and commercial potential. Although different approaches have been proposed to tackle this problem, it still remains challenging due to factors like abrupt appearance changes and severe object occlusions. In this work, we contribute the first comprehensive and most recent review on this problem. We inspect the recent advances in various aspects and propose some interesting directions for future research. To the best of our knowledge, there has not been any extensive review on this topic in the community. We endeavor to provide a thorough review on the development of this problem in recent decades. The main contributions of this review are fourfold: 1) Key aspects in an MOT system, including formulation, categorization, key principles, evaluation of MOT are discussed; 2) Instead of enumerating individual works, we discuss existing approaches according to various aspects, in each of which methods are divided into different groups and each group is discussed in detail for the principles, advances and drawbacks; 3) We examine experiments of existing publications and summarize results on popular datasets to provide quantitative and comprehensive comparisons. By analyzing the results from different perspectives, we have verified some basic agreements in the field; and 4) We provide a discussion about issues of MOT research, as well as some interesting directions which will become potential research effort in the future.}
}


@inproceedings{Bewley2016,
abstract = {This paper explores a pragmatic approach to multiple object tracking where the main focus is to associate objects efficiently for online and realtime applications. To this end, detection quality is identified as a key factor influencing tracking performance, where changing the detector can improve tracking by up to 18.9%. Despite only using a rudimentary combination of familiar techniques such as the Kalman Filter and Hungarian algorithm for the tracking components, this approach achieves an accuracy comparable to state-of-the-art online trackers. Furthermore, due to the simplicity of our tracking method, the tracker updates at a rate of 260 Hz which is over 20x faster than other state-of-the-art trackers.},
archivePrefix = {arXiv},
arxivId = {1602.00763},
author = {Bewley, Alex and Ge, Zongyuan and Ott, Lionel and Ramos, Fabio and Upcroft, Ben},
booktitle = {Proceedings - International Conference on Image Processing, ICIP},
doi = {10.1109/ICIP.2016.7533003},
eprint = {1602.00763},
isbn = {9781467399616},
issn = {15224880},
keywords = {Computer Vision,Data Association,Detection,Multiple Object Tracking},
pages = {3464--3468},
title = {{Simple online and realtime tracking}},
url = {https://github.com/abewley/sort},
volume = {2016-Augus},
year = {2016}
}


@inproceedings{Caesar2020,
abstract = {Robust detection and tracking of objects is crucial for the deployment of autonomous vehicle technology. Image based benchmark datasets have driven development in computer vision tasks such as object detection, tracking and segmentation of agents in the environment. Most autonomous vehicles, however, carry a combination of cameras and range sensors such as lidar and radar. As machine learning based methods for detection and tracking become more prevalent, there is a need to train and evaluate such methods on datasets containing range sensor data along with images. In this work we present nuTonomy scenes (nuScenes), the first dataset to carry the full autonomous vehicle sensor suite: 6 cameras, 5 radars and 1 lidar, all with full 360 degree field of view. nuScenes comprises 1000 scenes, each 20s long and fully annotated with 3D bounding boxes for 23 classes and 8 attributes. It has 7x as many annotations and 100x as many images as the pioneering KITTI dataset. We define novel 3D detection and tracking metrics. We also provide careful dataset analysis as well as baselines for lidar and image based detection and tracking. Data, development kit and more information are available online1},
archivePrefix = {arXiv},
arxivId = {1903.11027},
author = {Caesar, Holger and Bankiti, Varun and Lang, Alex H and Vora, Sourabh and Liong, Venice Erin and Xu, Qiang and Krishnan, Anush and Pan, Yu and Baldan, Giancarlo and Beijbom, Oscar},
booktitle = {Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition},
doi = {10.1109/CVPR42600.2020.01164},
eprint = {1903.11027},
file = {:C\:/Users/mchag/AppData/Local/Mendeley Ltd./Mendeley Desktop/Downloaded/Caesar et al. - Unknown - nuScenes A multimodal dataset for autonomous driving.pdf:pdf},
issn = {10636919},
mendeley-groups = {PhD/multi_object_tracking,PhD/multi_object_tracking/datasets_benchmarks},
pages = {11618--11628},
title = {{Nuscenes: A multimodal dataset for autonomous driving}},
year = {2020}
}

@article{Dendorfer2020,
archivePrefix = {arXiv},
arxivId = {2003.09003},
author = {Dendorfer, Patrick and Rezatofighi, Hamid and Milan, Anton and Shi, Javen and Cremers, Daniel and Reid, Ian and Roth, Stefan and Schindler, Konrad and Leal-Taix{\'{e}}, Laura},
eprint = {2003.09003},
keywords = {Index Terms-multiple people tracking,benchmark,dataset !,evaluation metrics},
title = {{MOT20: A benchmark for multi object tracking in crowded scenes}},
year = {2020},
journal = {},
}


@inproceedings{Wojke2018,
abstract = {Simple Online and Realtime Tracking (SORT) is a pragmatic approach to multiple object tracking with a focus on simple, effective algorithms. In this paper, we integrate appearance information to improve the performance of SORT. Due to this extension we are able to track objects through longer periods of occlusions, effectively reducing the number of identity switches. In spirit of the original framework we place much of the computational complexity into an offline pre-training stage where we learn a deep association metric on a largescale person re-identification dataset. During online application, we establish measurement-to-track associations using nearest neighbor queries in visual appearance space. Experimental evaluation shows that our extensions reduce the number of identity switches by 45%, achieving overall competitive performance at high frame rates.},
archivePrefix = {arXiv},
arxivId = {1703.07402},
author = {Wojke, Nicolai and Bewley, Alex and Paulus, Dietrich},
booktitle = {Proceedings - International Conference on Image Processing, ICIP},
doi = {10.1109/ICIP.2017.8296962},
eprint = {1703.07402},
file = {:C\:/Users/mchag/AppData/Local/Mendeley Ltd./Mendeley Desktop/Downloaded/Wojke, Bewley, Paulus - Unknown - SIMPLE ONLINE AND REALTIME TRACKING WITH A DEEP ASSOCIATION METRIC.pdf:pdf},
isbn = {9781509021758},
issn = {15224880},
keywords = {Computer Vision,Data Association,Multiple Object Tracking},
pages = {3645--3649},
title = {{Simple online and realtime tracking with a deep association metric}},
volume = {2017-Septe},
year = {2018}
}

@article{Chu2021,
abstract = {Tracking multiple objects in videos relies on modeling the spatial-temporal interactions of the objects. In this paper, we propose a solution named TransMOT, which leverages powerful graph transformers to efficiently model the spatial and temporal interactions among the objects. TransMOT effectively models the interactions of a large number of objects by arranging the trajectories of the tracked objects as a set of sparse weighted graphs, and constructing a spatial graph transformer encoder layer, a temporal transformer encoder layer, and a spatial graph transformer decoder layer based on the graphs. TransMOT is not only more computationally efficient than the traditional Transformer, but it also achieves better tracking accuracy. To further improve the tracking speed and accuracy, we propose a cascade association framework to handle low-score detections and long-term occlusions that require large computational resources to model in TransMOT. The proposed method is evaluated on multiple benchmark datasets including MOT15, MOT16, MOT17, and MOT20, and it achieves state-of-the-art performance on all the datasets.},
archivePrefix = {arXiv},
arxivId = {2104.00194},
author = {Chu, Peng and Wang, Jiang and You, Quanzeng and Ling, Haibin and Liu, Zicheng},
eprint = {2104.00194},
file = {:C\:/Users/mchag/AppData/Local/Mendeley Ltd./Mendeley Desktop/Downloaded/Chu et al. - Unknown - TransMOT Spatial-Temporal Graph Transformer for Multiple Object Tracking.pdf:pdf},
title = {{TransMOT: Spatial-Temporal Graph Transformer for Multiple Object Tracking}},
url = {http://arxiv.org/abs/2104.00194},
year = {2021},
journal = {}
}


@article{Wang2021,
author = {Yongxin Wang and Kris Kitani and Xinshuo Weng},
title = {Joint Object Detection and Multi-Object Tracking with Graph Neural Networks},
booktitle = {Proceedings of (ICRA) International Conference on Robotics and Automation},
year = {2021},
month = {May},
journal = {}
}

@article{bernardin2008,
author = {Bernardin, Keni and Stiefelhagen, Rainer},
year = {2008},
month = {01},
pages = {},
title = {Evaluating multiple object tracking performance: The CLEAR MOT metrics},
volume = {2008},
journal = {EURASIP Journal on Image and Video Processing},
doi = {10.1155/2008/246309}
}

@inproceedings{RistaniSZCT16,
  title={Performance measures and a data set for multi-target, multi-camera tracking},
  author={Ristani, Ergys and Solera, Francesco and Zou, Roger and Cucchiara, Rita and Tomasi, Carlo},
  booktitle={European conference on computer vision},
  pages={17--35},
  year={2016},
  organization={Springer}
}