Evolutionary Feature Learning for 3D Object Recognition

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Three-dimensional (3D) object recognition is a challenging task for many applications including autonomous robot navigation and scene understanding. Accurate recognition relies on the selection/learning of discriminative features that are in turn used to uniquely characterize the objects. This paper proposes a novel Evolutionary Feature Learning (EFL) technique for 3D object recognition. The proposed novel automatic feature learning approach can operate directly on 3D raw data, alleviating the need for data pre-processing, human expertise and/or defining a large set of parameters. EFL offers smart search strategy to learn the best features in a huge feature space to achieve superior recognition performance. The proposed technique has been extensively evaluated for the task of 3D object recognition on four popular datasets including Washington RGB-D (low resolution 3D Video), CIN 2D3D, Willow 2D3D and ETH-80 object dataset. Reported experimental results and evaluation against existing state-of-the-art methods (e.g. unsupervised dictionary learning and deep networks) show that the proposed EFL consistently achieves superior performance on all these datasets.

Original languageEnglish
Pages (from-to)2434-2444
JournalIEEE Access
Volume6
DOIs
Publication statusPublished - 2018

Fingerprint

Object recognition
Salix
Glossaries
Navigation
Robots
Processing

Cite this

@article{8400e28b619442c58da470e5e5571dd7,
title = "Evolutionary Feature Learning for 3D Object Recognition",
abstract = "Three-dimensional (3D) object recognition is a challenging task for many applications including autonomous robot navigation and scene understanding. Accurate recognition relies on the selection/learning of discriminative features that are in turn used to uniquely characterize the objects. This paper proposes a novel Evolutionary Feature Learning (EFL) technique for 3D object recognition. The proposed novel automatic feature learning approach can operate directly on 3D raw data, alleviating the need for data pre-processing, human expertise and/or defining a large set of parameters. EFL offers smart search strategy to learn the best features in a huge feature space to achieve superior recognition performance. The proposed technique has been extensively evaluated for the task of 3D object recognition on four popular datasets including Washington RGB-D (low resolution 3D Video), CIN 2D3D, Willow 2D3D and ETH-80 object dataset. Reported experimental results and evaluation against existing state-of-the-art methods (e.g. unsupervised dictionary learning and deep networks) show that the proposed EFL consistently achieves superior performance on all these datasets.",
keywords = "3D Object Recognition, Evolutionary Algorithms, Feature learning",
author = "Shah, {Syed Afaq Ali} and Mohammed Bennamoun and Farid Boussaid and Lyndon While",
year = "2018",
doi = "10.1109/ACCESS.2017.2783331",
language = "English",
volume = "6",
pages = "2434--2444",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",

}

TY - JOUR

T1 - Evolutionary Feature Learning for 3D Object Recognition

AU - Shah, Syed Afaq Ali

AU - Bennamoun, Mohammed

AU - Boussaid, Farid

AU - While, Lyndon

PY - 2018

Y1 - 2018

N2 - Three-dimensional (3D) object recognition is a challenging task for many applications including autonomous robot navigation and scene understanding. Accurate recognition relies on the selection/learning of discriminative features that are in turn used to uniquely characterize the objects. This paper proposes a novel Evolutionary Feature Learning (EFL) technique for 3D object recognition. The proposed novel automatic feature learning approach can operate directly on 3D raw data, alleviating the need for data pre-processing, human expertise and/or defining a large set of parameters. EFL offers smart search strategy to learn the best features in a huge feature space to achieve superior recognition performance. The proposed technique has been extensively evaluated for the task of 3D object recognition on four popular datasets including Washington RGB-D (low resolution 3D Video), CIN 2D3D, Willow 2D3D and ETH-80 object dataset. Reported experimental results and evaluation against existing state-of-the-art methods (e.g. unsupervised dictionary learning and deep networks) show that the proposed EFL consistently achieves superior performance on all these datasets.

AB - Three-dimensional (3D) object recognition is a challenging task for many applications including autonomous robot navigation and scene understanding. Accurate recognition relies on the selection/learning of discriminative features that are in turn used to uniquely characterize the objects. This paper proposes a novel Evolutionary Feature Learning (EFL) technique for 3D object recognition. The proposed novel automatic feature learning approach can operate directly on 3D raw data, alleviating the need for data pre-processing, human expertise and/or defining a large set of parameters. EFL offers smart search strategy to learn the best features in a huge feature space to achieve superior recognition performance. The proposed technique has been extensively evaluated for the task of 3D object recognition on four popular datasets including Washington RGB-D (low resolution 3D Video), CIN 2D3D, Willow 2D3D and ETH-80 object dataset. Reported experimental results and evaluation against existing state-of-the-art methods (e.g. unsupervised dictionary learning and deep networks) show that the proposed EFL consistently achieves superior performance on all these datasets.

KW - 3D Object Recognition

KW - Evolutionary Algorithms

KW - Feature learning

UR - http://www.scopus.com/inward/record.url?scp=85038834528&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2017.2783331

DO - 10.1109/ACCESS.2017.2783331

M3 - Article

VL - 6

SP - 2434

EP - 2444

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

ER -