Automatic Training Data Generation Method for Pixel-Level Road Lane Segmentation

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Lane detection or road detection is one of the key features of autonomous driving. By using deep convolutional neural network based semantic segmentation, we can build models with high accuracy and robustness. However, training a pixel-level semantic segmentation needs very fine-labeled training data, which requires large amount of labor. In this paper, we propose an automatic training data generating method, which can significantly reduce the effort of the training phase. Experiments prove that our method can generate high-quality training data for lane segmentation task.

Original languageEnglish
Title of host publicationGenetic and Evolutionary Computing - Proceedings of the 12th International Conference on Genetic and Evolutionary Computing, 2018
EditorsShih-Pang Tseng, Bixia Sui, Jeng-Shyang Pan, Jerry Chun-Wei Lin
PublisherSpringer-Verlag
Pages473-481
Number of pages9
ISBN (Print)9789811358401
DOIs
Publication statusPublished - 2019 Jan 1
Event12th International Conference on Genetic and Evolutionary Computing, ICGEC 2018 - Changzhou, China
Duration: 2018 Dec 142018 Dec 17

Publication series

NameAdvances in Intelligent Systems and Computing
Volume834
ISSN (Print)2194-5357

Conference

Conference12th International Conference on Genetic and Evolutionary Computing, ICGEC 2018
CountryChina
CityChangzhou
Period18/12/1418/12/17

Fingerprint

Pixels
Semantics
Personnel
Neural networks
Experiments

Keywords

  • Lane detection
  • Semantic segmentation
  • Training data generation

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Computer Science(all)

Cite this

Pan, X., Wu, Y., & Ogai, H. (2019). Automatic Training Data Generation Method for Pixel-Level Road Lane Segmentation. In S-P. Tseng, B. Sui, J-S. Pan, & J. C-W. Lin (Eds.), Genetic and Evolutionary Computing - Proceedings of the 12th International Conference on Genetic and Evolutionary Computing, 2018 (pp. 473-481). (Advances in Intelligent Systems and Computing; Vol. 834). Springer-Verlag. https://doi.org/10.1007/978-981-13-5841-8_49

Automatic Training Data Generation Method for Pixel-Level Road Lane Segmentation. / Pan, Xun; Wu, Yutian; Ogai, Harutoshi.

Genetic and Evolutionary Computing - Proceedings of the 12th International Conference on Genetic and Evolutionary Computing, 2018. ed. / Shih-Pang Tseng; Bixia Sui; Jeng-Shyang Pan; Jerry Chun-Wei Lin. Springer-Verlag, 2019. p. 473-481 (Advances in Intelligent Systems and Computing; Vol. 834).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Pan, X, Wu, Y & Ogai, H 2019, Automatic Training Data Generation Method for Pixel-Level Road Lane Segmentation. in S-P Tseng, B Sui, J-S Pan & JC-W Lin (eds), Genetic and Evolutionary Computing - Proceedings of the 12th International Conference on Genetic and Evolutionary Computing, 2018. Advances in Intelligent Systems and Computing, vol. 834, Springer-Verlag, pp. 473-481, 12th International Conference on Genetic and Evolutionary Computing, ICGEC 2018, Changzhou, China, 18/12/14. https://doi.org/10.1007/978-981-13-5841-8_49
Pan X, Wu Y, Ogai H. Automatic Training Data Generation Method for Pixel-Level Road Lane Segmentation. In Tseng S-P, Sui B, Pan J-S, Lin JC-W, editors, Genetic and Evolutionary Computing - Proceedings of the 12th International Conference on Genetic and Evolutionary Computing, 2018. Springer-Verlag. 2019. p. 473-481. (Advances in Intelligent Systems and Computing). https://doi.org/10.1007/978-981-13-5841-8_49
Pan, Xun ; Wu, Yutian ; Ogai, Harutoshi. / Automatic Training Data Generation Method for Pixel-Level Road Lane Segmentation. Genetic and Evolutionary Computing - Proceedings of the 12th International Conference on Genetic and Evolutionary Computing, 2018. editor / Shih-Pang Tseng ; Bixia Sui ; Jeng-Shyang Pan ; Jerry Chun-Wei Lin. Springer-Verlag, 2019. pp. 473-481 (Advances in Intelligent Systems and Computing).
@inproceedings{6df01ded9a52474caa6adb98c1311b26,
title = "Automatic Training Data Generation Method for Pixel-Level Road Lane Segmentation",
abstract = "Lane detection or road detection is one of the key features of autonomous driving. By using deep convolutional neural network based semantic segmentation, we can build models with high accuracy and robustness. However, training a pixel-level semantic segmentation needs very fine-labeled training data, which requires large amount of labor. In this paper, we propose an automatic training data generating method, which can significantly reduce the effort of the training phase. Experiments prove that our method can generate high-quality training data for lane segmentation task.",
keywords = "Lane detection, Semantic segmentation, Training data generation",
author = "Xun Pan and Yutian Wu and Harutoshi Ogai",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-981-13-5841-8_49",
language = "English",
isbn = "9789811358401",
series = "Advances in Intelligent Systems and Computing",
publisher = "Springer-Verlag",
pages = "473--481",
editor = "Shih-Pang Tseng and Bixia Sui and Jeng-Shyang Pan and Lin, {Jerry Chun-Wei}",
booktitle = "Genetic and Evolutionary Computing - Proceedings of the 12th International Conference on Genetic and Evolutionary Computing, 2018",

}

TY - GEN

T1 - Automatic Training Data Generation Method for Pixel-Level Road Lane Segmentation

AU - Pan, Xun

AU - Wu, Yutian

AU - Ogai, Harutoshi

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Lane detection or road detection is one of the key features of autonomous driving. By using deep convolutional neural network based semantic segmentation, we can build models with high accuracy and robustness. However, training a pixel-level semantic segmentation needs very fine-labeled training data, which requires large amount of labor. In this paper, we propose an automatic training data generating method, which can significantly reduce the effort of the training phase. Experiments prove that our method can generate high-quality training data for lane segmentation task.

AB - Lane detection or road detection is one of the key features of autonomous driving. By using deep convolutional neural network based semantic segmentation, we can build models with high accuracy and robustness. However, training a pixel-level semantic segmentation needs very fine-labeled training data, which requires large amount of labor. In this paper, we propose an automatic training data generating method, which can significantly reduce the effort of the training phase. Experiments prove that our method can generate high-quality training data for lane segmentation task.

KW - Lane detection

KW - Semantic segmentation

KW - Training data generation

UR - http://www.scopus.com/inward/record.url?scp=85066033114&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066033114&partnerID=8YFLogxK

U2 - 10.1007/978-981-13-5841-8_49

DO - 10.1007/978-981-13-5841-8_49

M3 - Conference contribution

AN - SCOPUS:85066033114

SN - 9789811358401

T3 - Advances in Intelligent Systems and Computing

SP - 473

EP - 481

BT - Genetic and Evolutionary Computing - Proceedings of the 12th International Conference on Genetic and Evolutionary Computing, 2018

A2 - Tseng, Shih-Pang

A2 - Sui, Bixia

A2 - Pan, Jeng-Shyang

A2 - Lin, Jerry Chun-Wei

PB - Springer-Verlag

ER -