Deep learning-based 3D local feature descriptor from Mercator projections

Masoumeh Rezaei, Mehdi Rezaeian, Vali Derhami, Ferdous Sohel, Mohammed Bennamoun

Research output: Contribution to journalArticle

Abstract

Point clouds provide rich geometric information about a shape and a deep neural network can be used to learn effective and robust features. In this paper, we propose a novel local feature descriptor, which employs a Siamese network to directly learn robust features from the point clouds. We use a data representation based on the Mercator projection, then we use a Siamese network to map this projection into a 32-dimensional local descriptor. To validate the proposed method, we have compared it with seven state-of-the-art descriptor methods. Experimental results show the superiority of the proposed method compared to existing methods in terms of descriptiveness and robustness against noise and varying mesh resolutions.
Original languageEnglish
Article number101771
JournalComputer Aided Geometric Design
Volume74
DOIs
Publication statusPublished - Oct 2019

Fingerprint

Local Features
Descriptors
Projection
Point Cloud
Noise Robustness
Mesh
Neural Networks
Learning
Deep learning
Deep neural networks
Experimental Results

Cite this

@article{975e5cb31b6c497d87995f331f26a461,
title = "Deep learning-based 3D local feature descriptor from Mercator projections",
abstract = "Point clouds provide rich geometric information about a shape and a deep neural network can be used to learn effective and robust features. In this paper, we propose a novel local feature descriptor, which employs a Siamese network to directly learn robust features from the point clouds. We use a data representation based on the Mercator projection, then we use a Siamese network to map this projection into a 32-dimensional local descriptor. To validate the proposed method, we have compared it with seven state-of-the-art descriptor methods. Experimental results show the superiority of the proposed method compared to existing methods in terms of descriptiveness and robustness against noise and varying mesh resolutions.",
author = "Masoumeh Rezaei and Mehdi Rezaeian and Vali Derhami and Ferdous Sohel and Mohammed Bennamoun",
year = "2019",
month = "10",
doi = "10.1016/j.cagd.2019.101771",
language = "English",
volume = "74",
journal = "Computer Aided Geometric Design",
issn = "0167-8396",
publisher = "Elsevier",

}

Deep learning-based 3D local feature descriptor from Mercator projections. / Rezaei, Masoumeh; Rezaeian, Mehdi; Derhami, Vali; Sohel, Ferdous; Bennamoun, Mohammed.

In: Computer Aided Geometric Design, Vol. 74, 101771, 10.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Deep learning-based 3D local feature descriptor from Mercator projections

AU - Rezaei, Masoumeh

AU - Rezaeian, Mehdi

AU - Derhami, Vali

AU - Sohel, Ferdous

AU - Bennamoun, Mohammed

PY - 2019/10

Y1 - 2019/10

N2 - Point clouds provide rich geometric information about a shape and a deep neural network can be used to learn effective and robust features. In this paper, we propose a novel local feature descriptor, which employs a Siamese network to directly learn robust features from the point clouds. We use a data representation based on the Mercator projection, then we use a Siamese network to map this projection into a 32-dimensional local descriptor. To validate the proposed method, we have compared it with seven state-of-the-art descriptor methods. Experimental results show the superiority of the proposed method compared to existing methods in terms of descriptiveness and robustness against noise and varying mesh resolutions.

AB - Point clouds provide rich geometric information about a shape and a deep neural network can be used to learn effective and robust features. In this paper, we propose a novel local feature descriptor, which employs a Siamese network to directly learn robust features from the point clouds. We use a data representation based on the Mercator projection, then we use a Siamese network to map this projection into a 32-dimensional local descriptor. To validate the proposed method, we have compared it with seven state-of-the-art descriptor methods. Experimental results show the superiority of the proposed method compared to existing methods in terms of descriptiveness and robustness against noise and varying mesh resolutions.

U2 - 10.1016/j.cagd.2019.101771

DO - 10.1016/j.cagd.2019.101771

M3 - Article

VL - 74

JO - Computer Aided Geometric Design

JF - Computer Aided Geometric Design

SN - 0167-8396

M1 - 101771

ER -