Adaptive Update in Deep Learning Algorithms for LiDAR Data Semantic Segmentation

Nur Hamid; Ari Wibisono; Ahmad Gamal; Ronni Ardhianto; Wisnu Jatmiko

doi:10.1109/TENSYMP50017.2020.9230926

Adaptive Update in Deep Learning Algorithms for LiDAR Data Semantic Segmentation

Nur Hamid
, Ari Wibisono
, Ahmad Gamal
, Ronni Ardhianto
, Wisnu Jatmiko

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

2 Scopus citations

Abstract

LiDAR data widely replaces 2-dimensional data for geographic data representation because of its information complexity. One of the LiDAR data processing tasks is semantic segmentation which has been developed by deep learning models. These algorithms use Euclidean distance representation to express the distance between the points, whereas LiDAR data with random properties are not suitable to use this distance representation. Therefore, this study proposes the non-Euclidean distance representation which is adaptively updated using their covariance values. The proposed method results the accuracy of 75.55%, better than the baseline PointNet of 65.08% and Dynamic Graph CNN of 72.56% with the dataset from the author. This performance improvement is because of multiplication with the inverse covariance value of point cloud data increasing the points similarity to the class.

Original language	English
Title of host publication	2020 IEEE Region 10 Symposium, TENSYMP 2020
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	1038-1041
Number of pages	4
ISBN (Electronic)	9781728173665
DOIs	https://doi.org/10.1109/TENSYMP50017.2020.9230926
State	Published - 5 Jun 2020
Externally published	Yes

Publication series

Name	2020 IEEE Region 10 Symposium, TENSYMP 2020

Bibliographical note

Publisher Copyright:
© 2020 IEEE.

Keywords

LiDAR data
deep learning
graph convolutional network
land cover semantic segmentation
non-Euclidean

ASJC Scopus subject areas

Artificial Intelligence
Computer Networks and Communications
Computer Vision and Pattern Recognition
Signal Processing
Energy Engineering and Power Technology
Biomedical Engineering
Electrical and Electronic Engineering
Health Informatics

Access to Document

10.1109/TENSYMP50017.2020.9230926

Cite this

Hamid, N., Wibisono, A., Gamal, A., Ardhianto, R., & Jatmiko, W. (2020). Adaptive Update in Deep Learning Algorithms for LiDAR Data Semantic Segmentation. In 2020 IEEE Region 10 Symposium, TENSYMP 2020 (pp. 1038-1041). Article 9230926 (2020 IEEE Region 10 Symposium, TENSYMP 2020). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/TENSYMP50017.2020.9230926

@inproceedings{0f3fafa20de042c48846466ae38bb972,

title = "Adaptive Update in Deep Learning Algorithms for LiDAR Data Semantic Segmentation",

abstract = "LiDAR data widely replaces 2-dimensional data for geographic data representation because of its information complexity. One of the LiDAR data processing tasks is semantic segmentation which has been developed by deep learning models. These algorithms use Euclidean distance representation to express the distance between the points, whereas LiDAR data with random properties are not suitable to use this distance representation. Therefore, this study proposes the non-Euclidean distance representation which is adaptively updated using their covariance values. The proposed method results the accuracy of 75.55\%, better than the baseline PointNet of 65.08\% and Dynamic Graph CNN of 72.56\% with the dataset from the author. This performance improvement is because of multiplication with the inverse covariance value of point cloud data increasing the points similarity to the class.",

keywords = "LiDAR data, deep learning, graph convolutional network, land cover semantic segmentation, non-Euclidean",

author = "Nur Hamid and Ari Wibisono and Ahmad Gamal and Ronni Ardhianto and Wisnu Jatmiko",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.",

year = "2020",

month = jun,

day = "5",

doi = "10.1109/TENSYMP50017.2020.9230926",

language = "English",

series = "2020 IEEE Region 10 Symposium, TENSYMP 2020",

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

pages = "1038--1041",

booktitle = "2020 IEEE Region 10 Symposium, TENSYMP 2020",

address = "United States",

}

Hamid, N, Wibisono, A, Gamal, A, Ardhianto, R & Jatmiko, W 2020, Adaptive Update in Deep Learning Algorithms for LiDAR Data Semantic Segmentation. in 2020 IEEE Region 10 Symposium, TENSYMP 2020., 9230926, 2020 IEEE Region 10 Symposium, TENSYMP 2020, Institute of Electrical and Electronics Engineers Inc., pp. 1038-1041. https://doi.org/10.1109/TENSYMP50017.2020.9230926

Adaptive Update in Deep Learning Algorithms for LiDAR Data Semantic Segmentation. / Hamid, Nur; Wibisono, Ari; Gamal, Ahmad et al.
2020 IEEE Region 10 Symposium, TENSYMP 2020. Institute of Electrical and Electronics Engineers Inc., 2020. p. 1038-1041 9230926 (2020 IEEE Region 10 Symposium, TENSYMP 2020).

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

TY - GEN

T1 - Adaptive Update in Deep Learning Algorithms for LiDAR Data Semantic Segmentation

AU - Hamid, Nur

AU - Wibisono, Ari

AU - Gamal, Ahmad

AU - Ardhianto, Ronni

AU - Jatmiko, Wisnu

PY - 2020/6/5

Y1 - 2020/6/5

N2 - LiDAR data widely replaces 2-dimensional data for geographic data representation because of its information complexity. One of the LiDAR data processing tasks is semantic segmentation which has been developed by deep learning models. These algorithms use Euclidean distance representation to express the distance between the points, whereas LiDAR data with random properties are not suitable to use this distance representation. Therefore, this study proposes the non-Euclidean distance representation which is adaptively updated using their covariance values. The proposed method results the accuracy of 75.55%, better than the baseline PointNet of 65.08% and Dynamic Graph CNN of 72.56% with the dataset from the author. This performance improvement is because of multiplication with the inverse covariance value of point cloud data increasing the points similarity to the class.

AB - LiDAR data widely replaces 2-dimensional data for geographic data representation because of its information complexity. One of the LiDAR data processing tasks is semantic segmentation which has been developed by deep learning models. These algorithms use Euclidean distance representation to express the distance between the points, whereas LiDAR data with random properties are not suitable to use this distance representation. Therefore, this study proposes the non-Euclidean distance representation which is adaptively updated using their covariance values. The proposed method results the accuracy of 75.55%, better than the baseline PointNet of 65.08% and Dynamic Graph CNN of 72.56% with the dataset from the author. This performance improvement is because of multiplication with the inverse covariance value of point cloud data increasing the points similarity to the class.

KW - LiDAR data

KW - deep learning

KW - graph convolutional network

KW - land cover semantic segmentation

KW - non-Euclidean

UR - https://www.scopus.com/pages/publications/85096411568

U2 - 10.1109/TENSYMP50017.2020.9230926

DO - 10.1109/TENSYMP50017.2020.9230926

M3 - Conference contribution

AN - SCOPUS:85096411568

T3 - 2020 IEEE Region 10 Symposium, TENSYMP 2020

SP - 1038

EP - 1041

BT - 2020 IEEE Region 10 Symposium, TENSYMP 2020

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Adaptive Update in Deep Learning Algorithms for LiDAR Data Semantic Segmentation

Abstract

Publication series

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this