Semi-automatic training of a vehicle make and model recognition system

M.H. Zwemer; G.M.Y.E. Brouwers; R.G.J. Wijnhoven; P.H.N. de With

doi:10.1007/978-3-319-68548-9_30

Semi-automatic training of a vehicle make and model recognition system

M.H. Zwemer, G.M.Y.E. Brouwers, R.G.J. Wijnhoven, P.H.N. de With

Video Coding & Architectures

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

2 Citations (Scopus)

1 Downloads (Pure)

Abstract

We propose a system for vehicle Make and Model Recognition (MMR) that automatically detects and classifies the make and model from a live camera mounted above the highway. Our system consists of a vehicle detection and MMR classification component. The vehicle detector is based on HOG features and can locate 98%
98%
of the vehicles with minimum false detections. We use a Convolutional Neural Network (CNN) for MMR classification on the vehicle locations. We propose a semi-automatic data-selection approach for the vehicle detector and the MMR classifier, by using an Automatic Number Plate Recognition engine for annotating new images, requiring minimal human annotation effort. In our results we show that our MMR classification has a top-1 accuracy of 98%
98%
for 500 vehicle models, where more than 500 training samples per model are desired to obtain accurate classification.

Original language	English
Title of host publication	Image Analysis and Processing - ICIAP 2017
Subtitle of host publication	19th International Conference, Catania, Italy, September 11-15, 2017, Proceedings, Part II
Editors	Sebastiano Battiato, Giovanni Gallo, Filippo Stanco, Raimondo Schettini
Place of Publication	Cham
Publisher	Springer
Pages	321-332
Number of pages	12
ISBN (Electronic)	978-3-319-68548-9
ISBN (Print)	978-3-319-68547-2
DOIs	https://doi.org/10.1007/978-3-319-68548-9_30
Publication status	Published - 2017
Event	19th International Conference on Image Analysis and Processing (ICIAP 2017) - Catania, Italy Duration: 11 Sept 2017 → 15 Sept 2017 Conference number: 19 http://www.iciap2017.com/

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10485 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	19th International Conference on Image Analysis and Processing (ICIAP 2017)
Abbreviated title	ICIAP 2017
Country/Territory	Italy
City	Catania
Period	11/09/17 → 15/09/17
Internet address	http://www.iciap2017.com/

Access to Document

10.1007/978-3-319-68548-9_30

Cite this

Zwemer, M. H., Brouwers, G. M. Y. E., Wijnhoven, R. G. J., & de With, P. H. N. (2017). Semi-automatic training of a vehicle make and model recognition system. In S. Battiato, G. Gallo, F. Stanco, & R. Schettini (Eds.), Image Analysis and Processing - ICIAP 2017: 19th International Conference, Catania, Italy, September 11-15, 2017, Proceedings, Part II (pp. 321-332). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10485 LNCS). Springer. https://doi.org/10.1007/978-3-319-68548-9_30

Zwemer, M.H. ; Brouwers, G.M.Y.E. ; Wijnhoven, R.G.J. et al. / Semi-automatic training of a vehicle make and model recognition system. Image Analysis and Processing - ICIAP 2017: 19th International Conference, Catania, Italy, September 11-15, 2017, Proceedings, Part II. editor / Sebastiano Battiato ; Giovanni Gallo ; Filippo Stanco ; Raimondo Schettini. Cham : Springer, 2017. pp. 321-332 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{d27f239c9fdb4b748a17f6dbe9392d11,

title = "Semi-automatic training of a vehicle make and model recognition system",

abstract = "We propose a system for vehicle Make and Model Recognition (MMR) that automatically detects and classifies the make and model from a live camera mounted above the highway. Our system consists of a vehicle detection and MMR classification component. The vehicle detector is based on HOG features and can locate 98% 98% of the vehicles with minimum false detections. We use a Convolutional Neural Network (CNN) for MMR classification on the vehicle locations. We propose a semi-automatic data-selection approach for the vehicle detector and the MMR classifier, by using an Automatic Number Plate Recognition engine for annotating new images, requiring minimal human annotation effort. In our results we show that our MMR classification has a top-1 accuracy of 98% 98% for 500 vehicle models, where more than 500 training samples per model are desired to obtain accurate classification.",

author = "M.H. Zwemer and G.M.Y.E. Brouwers and R.G.J. Wijnhoven and {de With}, P.H.N.",

year = "2017",

doi = "10.1007/978-3-319-68548-9_30",

language = "English",

isbn = "978-3-319-68547-2",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "321--332",

editor = "Sebastiano Battiato and Giovanni Gallo and Filippo Stanco and Raimondo Schettini",

booktitle = "Image Analysis and Processing - ICIAP 2017",

address = "Germany",

note = "19th International Conference on Image Analysis and Processing (ICIAP 2017), ICIAP 2017 ; Conference date: 11-09-2017 Through 15-09-2017",

url = "http://www.iciap2017.com/",

}

Zwemer, MH, Brouwers, GMYE, Wijnhoven, RGJ & de With, PHN 2017, Semi-automatic training of a vehicle make and model recognition system. in S Battiato, G Gallo, F Stanco & R Schettini (eds), Image Analysis and Processing - ICIAP 2017: 19th International Conference, Catania, Italy, September 11-15, 2017, Proceedings, Part II. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10485 LNCS, Springer, Cham, pp. 321-332, 19th International Conference on Image Analysis and Processing (ICIAP 2017), Catania, Italy, 11/09/17. https://doi.org/10.1007/978-3-319-68548-9_30

Semi-automatic training of a vehicle make and model recognition system. / Zwemer, M.H.; Brouwers, G.M.Y.E.; Wijnhoven, R.G.J. et al.
Image Analysis and Processing - ICIAP 2017: 19th International Conference, Catania, Italy, September 11-15, 2017, Proceedings, Part II. ed. / Sebastiano Battiato; Giovanni Gallo; Filippo Stanco; Raimondo Schettini. Cham: Springer, 2017. p. 321-332 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10485 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Semi-automatic training of a vehicle make and model recognition system

AU - Zwemer, M.H.

AU - Brouwers, G.M.Y.E.

AU - Wijnhoven, R.G.J.

AU - de With, P.H.N.

N1 - Conference code: 19

PY - 2017

Y1 - 2017

N2 - We propose a system for vehicle Make and Model Recognition (MMR) that automatically detects and classifies the make and model from a live camera mounted above the highway. Our system consists of a vehicle detection and MMR classification component. The vehicle detector is based on HOG features and can locate 98% 98% of the vehicles with minimum false detections. We use a Convolutional Neural Network (CNN) for MMR classification on the vehicle locations. We propose a semi-automatic data-selection approach for the vehicle detector and the MMR classifier, by using an Automatic Number Plate Recognition engine for annotating new images, requiring minimal human annotation effort. In our results we show that our MMR classification has a top-1 accuracy of 98% 98% for 500 vehicle models, where more than 500 training samples per model are desired to obtain accurate classification.

AB - We propose a system for vehicle Make and Model Recognition (MMR) that automatically detects and classifies the make and model from a live camera mounted above the highway. Our system consists of a vehicle detection and MMR classification component. The vehicle detector is based on HOG features and can locate 98% 98% of the vehicles with minimum false detections. We use a Convolutional Neural Network (CNN) for MMR classification on the vehicle locations. We propose a semi-automatic data-selection approach for the vehicle detector and the MMR classifier, by using an Automatic Number Plate Recognition engine for annotating new images, requiring minimal human annotation effort. In our results we show that our MMR classification has a top-1 accuracy of 98% 98% for 500 vehicle models, where more than 500 training samples per model are desired to obtain accurate classification.

UR - http://www.scopus.com/inward/record.url?scp=85032481869&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-68548-9_30

DO - 10.1007/978-3-319-68548-9_30

M3 - Conference contribution

SN - 978-3-319-68547-2

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 321

EP - 332

BT - Image Analysis and Processing - ICIAP 2017

A2 - Battiato, Sebastiano

A2 - Gallo, Giovanni

A2 - Stanco, Filippo

A2 - Schettini, Raimondo

PB - Springer

CY - Cham

T2 - 19th International Conference on Image Analysis and Processing (ICIAP 2017)

Y2 - 11 September 2017 through 15 September 2017

ER -

Zwemer MH, Brouwers GMYE, Wijnhoven RGJ, de With PHN. Semi-automatic training of a vehicle make and model recognition system. In Battiato S, Gallo G, Stanco F, Schettini R, editors, Image Analysis and Processing - ICIAP 2017: 19th International Conference, Catania, Italy, September 11-15, 2017, Proceedings, Part II. Cham: Springer. 2017. p. 321-332. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-68548-9_30

Semi-automatic training of a vehicle make and model recognition system

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this