2-D Deep Learning Model on 3-D Image Segmentation

Chien Chia Lee; Hsien I. Lin

doi:10.1109/ICMA49215.2020.9233871

2-D Deep Learning Model on 3-D Image Segmentation

Chien Chia Lee, Hsien I. Lin

Institute of Electrical and Control Engineering

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

1 Scopus citations

Abstract

For image segmentation of 3D objects, researchers recently focus on 3D image deep learning methods. These studies aim at developing robust and accurate deep learning models. However, 3D deep learning methods using point-cloud data are time-consuming. In this paper, we present a novel 3D image segmentation method based on a 2D deep learning model to achieve efficient segmentation performance. While using a single camera angle to collect object 3D information, we merely use the depth map and adopt a 2D deep-learning model to segment objects on the scene. We validated the proposed method by comparison with Point-Net. In our experiment, we provided an extensive comparison between Point-Net and our 2D deep-learning model. The result shows that our model had a similar accuracy but is much faster than Point-Net.

Original language	English
Title of host publication	2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	605-609
Number of pages	5
ISBN (Electronic)	9781728164151
DOIs	https://doi.org/10.1109/ICMA49215.2020.9233871
State	Published - 13 Oct 2020
Event	17th IEEE International Conference on Mechatronics and Automation, ICMA 2020 - Beijing, China Duration: 13 Oct 2020 → 16 Oct 2020

Publication series

Name	2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020

Conference

Conference	17th IEEE International Conference on Mechatronics and Automation, ICMA 2020
Country/Territory	China
City	Beijing
Period	13/10/20 → 16/10/20

Keywords

2D deep learning model
Image segmentation
point-cloud data
Point-Net

Access to Document

10.1109/ICMA49215.2020.9233871

Cite this

Lee, C. C., & Lin, H. I. (2020). 2-D Deep Learning Model on 3-D Image Segmentation. In 2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020 (pp. 605-609). Article 9233871 (2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICMA49215.2020.9233871

@inproceedings{06f0bd2058ea448a8ae7a4b0ff306072,

title = "2-D Deep Learning Model on 3-D Image Segmentation",

abstract = "For image segmentation of 3D objects, researchers recently focus on 3D image deep learning methods. These studies aim at developing robust and accurate deep learning models. However, 3D deep learning methods using point-cloud data are time-consuming. In this paper, we present a novel 3D image segmentation method based on a 2D deep learning model to achieve efficient segmentation performance. While using a single camera angle to collect object 3D information, we merely use the depth map and adopt a 2D deep-learning model to segment objects on the scene. We validated the proposed method by comparison with Point-Net. In our experiment, we provided an extensive comparison between Point-Net and our 2D deep-learning model. The result shows that our model had a similar accuracy but is much faster than Point-Net.",

keywords = "2D deep learning model, Image segmentation, point-cloud data, Point-Net",

author = "Lee, {Chien Chia} and Lin, {Hsien I.}",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 17th IEEE International Conference on Mechatronics and Automation, ICMA 2020 ; Conference date: 13-10-2020 Through 16-10-2020",

year = "2020",

month = oct,

day = "13",

doi = "10.1109/ICMA49215.2020.9233871",

language = "English",

series = "2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020",

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

pages = "605--609",

booktitle = "2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020",

address = "美國",

}

Lee, CC & Lin, HI 2020, 2-D Deep Learning Model on 3-D Image Segmentation. in 2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020., 9233871, 2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020, Institute of Electrical and Electronics Engineers Inc., pp. 605-609, 17th IEEE International Conference on Mechatronics and Automation, ICMA 2020, Beijing, China, 13/10/20. https://doi.org/10.1109/ICMA49215.2020.9233871

2-D Deep Learning Model on 3-D Image Segmentation. / Lee, Chien Chia; Lin, Hsien I.
2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020. Institute of Electrical and Electronics Engineers Inc., 2020. p. 605-609 9233871 (2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020).

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

TY - GEN

T1 - 2-D Deep Learning Model on 3-D Image Segmentation

AU - Lee, Chien Chia

AU - Lin, Hsien I.

PY - 2020/10/13

Y1 - 2020/10/13

N2 - For image segmentation of 3D objects, researchers recently focus on 3D image deep learning methods. These studies aim at developing robust and accurate deep learning models. However, 3D deep learning methods using point-cloud data are time-consuming. In this paper, we present a novel 3D image segmentation method based on a 2D deep learning model to achieve efficient segmentation performance. While using a single camera angle to collect object 3D information, we merely use the depth map and adopt a 2D deep-learning model to segment objects on the scene. We validated the proposed method by comparison with Point-Net. In our experiment, we provided an extensive comparison between Point-Net and our 2D deep-learning model. The result shows that our model had a similar accuracy but is much faster than Point-Net.

AB - For image segmentation of 3D objects, researchers recently focus on 3D image deep learning methods. These studies aim at developing robust and accurate deep learning models. However, 3D deep learning methods using point-cloud data are time-consuming. In this paper, we present a novel 3D image segmentation method based on a 2D deep learning model to achieve efficient segmentation performance. While using a single camera angle to collect object 3D information, we merely use the depth map and adopt a 2D deep-learning model to segment objects on the scene. We validated the proposed method by comparison with Point-Net. In our experiment, we provided an extensive comparison between Point-Net and our 2D deep-learning model. The result shows that our model had a similar accuracy but is much faster than Point-Net.

KW - 2D deep learning model

KW - Image segmentation

KW - point-cloud data

KW - Point-Net

UR - http://www.scopus.com/inward/record.url?scp=85096523092&partnerID=8YFLogxK

U2 - 10.1109/ICMA49215.2020.9233871

DO - 10.1109/ICMA49215.2020.9233871

M3 - Conference contribution

AN - SCOPUS:85096523092

T3 - 2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020

SP - 605

EP - 609

BT - 2020 IEEE International Conference on Mechatronics and Automation, ICMA 2020

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 17th IEEE International Conference on Mechatronics and Automation, ICMA 2020

Y2 - 13 October 2020 through 16 October 2020

ER -

2-D Deep Learning Model on 3-D Image Segmentation

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this