A novel simple light-weight neural network for road segmentation

Peng Wei Lin; Chih Ming Hsu

doi:10.1109/IS3C50286.2020.00101

A novel simple light-weight neural network for road segmentation

Peng Wei Lin
, Chih Ming Hsu

Department of Mechanical Engineering

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

Abstract

This study proposes simple methods to design a light-weight neural network. A deep learning domain has many state of the art neural networks so it is highly accurate for commonly used dataset such as ImageNet and Cifar-10, Cifar-100 and allows a rapid execution time and a small model. However, these state of the art neural networks are very complicated. This paper uses a VGG-16[1] model to reduce the size of the model and the inference time, but maintain accuracy. The semantic segmentation performance for the proposed method is compared to that for the VGG-16 model. The same full convolutional network (FCN) semantic segmentation algorithm [2] is used to compare the two models for the same semantic segmentation task. This study proposes an easier method to construct a light-weight neural network.

Original language	English
Title of host publication	Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	368-371
Number of pages	4
ISBN (Electronic)	9781728193625
DOIs	https://doi.org/10.1109/IS3C50286.2020.00101
State	Published - Nov 2020
Event	2020 International Symposium on Computer, Consumer and Control, IS3C 2020 - Taichung, Taiwan Duration: 13 Nov 2020 → 16 Nov 2020

Publication series

Name	Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020

Conference

Conference	2020 International Symposium on Computer, Consumer and Control, IS3C 2020
Country/Territory	Taiwan
City	Taichung
Period	13/11/20 → 16/11/20

Bibliographical note

Publisher Copyright:
© 2020 IEEE

Keywords

Light-weight
Neural network
Simple

Access to Document

10.1109/IS3C50286.2020.00101

Cite this

Lin, P. W., & Hsu, C. M. (2020). A novel simple light-weight neural network for road segmentation. In Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020 (pp. 368-371). Article 9394077 (Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IS3C50286.2020.00101

@inproceedings{db052a8dda0e4a52b6a23ad5314fa087,

title = "A novel simple light-weight neural network for road segmentation",

abstract = "This study proposes simple methods to design a light-weight neural network. A deep learning domain has many state of the art neural networks so it is highly accurate for commonly used dataset such as ImageNet and Cifar-10, Cifar-100 and allows a rapid execution time and a small model. However, these state of the art neural networks are very complicated. This paper uses a VGG-16[1] model to reduce the size of the model and the inference time, but maintain accuracy. The semantic segmentation performance for the proposed method is compared to that for the VGG-16 model. The same full convolutional network (FCN) semantic segmentation algorithm [2] is used to compare the two models for the same semantic segmentation task. This study proposes an easier method to construct a light-weight neural network.",

keywords = "Light-weight, Neural network, Simple",

author = "Lin, \{Peng Wei\} and Hsu, \{Chih Ming\}",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE; 2020 International Symposium on Computer, Consumer and Control, IS3C 2020 ; Conference date: 13-11-2020 Through 16-11-2020",

year = "2020",

month = nov,

doi = "10.1109/IS3C50286.2020.00101",

language = "???core.languages.en\_GB???",

series = "Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "368--371",

booktitle = "Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020",

}

Lin, PW & Hsu, CM 2020, A novel simple light-weight neural network for road segmentation. in Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020., 9394077, Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020, Institute of Electrical and Electronics Engineers Inc., pp. 368-371, 2020 International Symposium on Computer, Consumer and Control, IS3C 2020, Taichung, Taiwan, 13/11/20. https://doi.org/10.1109/IS3C50286.2020.00101

A novel simple light-weight neural network for road segmentation. / Lin, Peng Wei; Hsu, Chih Ming.
Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020. Institute of Electrical and Electronics Engineers Inc., 2020. p. 368-371 9394077 (Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020).

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

TY - GEN

T1 - A novel simple light-weight neural network for road segmentation

AU - Lin, Peng Wei

AU - Hsu, Chih Ming

PY - 2020/11

Y1 - 2020/11

N2 - This study proposes simple methods to design a light-weight neural network. A deep learning domain has many state of the art neural networks so it is highly accurate for commonly used dataset such as ImageNet and Cifar-10, Cifar-100 and allows a rapid execution time and a small model. However, these state of the art neural networks are very complicated. This paper uses a VGG-16[1] model to reduce the size of the model and the inference time, but maintain accuracy. The semantic segmentation performance for the proposed method is compared to that for the VGG-16 model. The same full convolutional network (FCN) semantic segmentation algorithm [2] is used to compare the two models for the same semantic segmentation task. This study proposes an easier method to construct a light-weight neural network.

AB - This study proposes simple methods to design a light-weight neural network. A deep learning domain has many state of the art neural networks so it is highly accurate for commonly used dataset such as ImageNet and Cifar-10, Cifar-100 and allows a rapid execution time and a small model. However, these state of the art neural networks are very complicated. This paper uses a VGG-16[1] model to reduce the size of the model and the inference time, but maintain accuracy. The semantic segmentation performance for the proposed method is compared to that for the VGG-16 model. The same full convolutional network (FCN) semantic segmentation algorithm [2] is used to compare the two models for the same semantic segmentation task. This study proposes an easier method to construct a light-weight neural network.

KW - Light-weight

KW - Neural network

KW - Simple

UR - https://www.scopus.com/pages/publications/85104869831

U2 - 10.1109/IS3C50286.2020.00101

DO - 10.1109/IS3C50286.2020.00101

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:85104869831

T3 - Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020

SP - 368

EP - 371

BT - Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020

Y2 - 13 November 2020 through 16 November 2020

ER -

Lin PW, Hsu CM. A novel simple light-weight neural network for road segmentation. In Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020. Institute of Electrical and Electronics Engineers Inc. 2020. p. 368-371. 9394077. (Proceedings - 2020 International Symposium on Computer, Consumer and Control, IS3C 2020). doi: 10.1109/IS3C50286.2020.00101

A novel simple light-weight neural network for road segmentation

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this