Distinction of 3D Objects and Scenes via Classification Network and Markov Random Field

Ran Song, Yonghuai Liu, Paul L Rosin

Research output: Contribution to journalArticle

Abstract

An importance measure of 3D objects inspired by human perception has a range of applications since people want computers to behave like humans in many tasks. This paper revisits a well-defined measure, distinction of 3D surface mesh, which indicates how important a region of a mesh is with respect to classification. We develop a method to compute it based on a classification network and a Markov Random Field (MRF). The classification network learns view-based distinction by handling multiple views of a 3D object. Using a classification network has an advantage of avoiding the training data problem which has become a major obstacle of applying deep learning to 3D object understanding tasks. The MRF estimates the parameters of a linear model for combining the view-based distinction maps. The experiments using several publicly accessible datasets show that the distinctive regions detected by our method are not just significantly different from those detected by methods based on handcrafted features, but more consistent with human perception. We also compare it with other perceptual measures and quantitatively evaluate its performance in the context of two applications. Furthermore, due to the view-based nature of our method, we are able to easily extend mesh distinction to 3D scenes containing multiple objects.
Original languageEnglish
JournalIEEE Transactions on Visualization and Computer Graphics
Early online date7 Dec 2018
DOIs
Publication statusE-pub ahead of print - 7 Dec 2018

Fingerprint

Experiments
Deep learning

Cite this

@article{bc4b7462db6f440da6bc93226b760fc7,
title = "Distinction of 3D Objects and Scenes via Classification Network and Markov Random Field",
abstract = "An importance measure of 3D objects inspired by human perception has a range of applications since people want computers to behave like humans in many tasks. This paper revisits a well-defined measure, distinction of 3D surface mesh, which indicates how important a region of a mesh is with respect to classification. We develop a method to compute it based on a classification network and a Markov Random Field (MRF). The classification network learns view-based distinction by handling multiple views of a 3D object. Using a classification network has an advantage of avoiding the training data problem which has become a major obstacle of applying deep learning to 3D object understanding tasks. The MRF estimates the parameters of a linear model for combining the view-based distinction maps. The experiments using several publicly accessible datasets show that the distinctive regions detected by our method are not just significantly different from those detected by methods based on handcrafted features, but more consistent with human perception. We also compare it with other perceptual measures and quantitatively evaluate its performance in the context of two applications. Furthermore, due to the view-based nature of our method, we are able to easily extend mesh distinction to 3D scenes containing multiple objects.",
author = "Ran Song and Yonghuai Liu and Rosin, {Paul L}",
year = "2018",
month = "12",
day = "7",
doi = "10.1109/TVCG.2018.2885750",
language = "English",
journal = "IEEE Transactions on Visualization and Computer Graphics",
issn = "1077-2626",
publisher = "Institute of Electrical and Electronics Engineers",

}

TY - JOUR

T1 - Distinction of 3D Objects and Scenes via Classification Network and Markov Random Field

AU - Song, Ran

AU - Liu, Yonghuai

AU - Rosin, Paul L

PY - 2018/12/7

Y1 - 2018/12/7

N2 - An importance measure of 3D objects inspired by human perception has a range of applications since people want computers to behave like humans in many tasks. This paper revisits a well-defined measure, distinction of 3D surface mesh, which indicates how important a region of a mesh is with respect to classification. We develop a method to compute it based on a classification network and a Markov Random Field (MRF). The classification network learns view-based distinction by handling multiple views of a 3D object. Using a classification network has an advantage of avoiding the training data problem which has become a major obstacle of applying deep learning to 3D object understanding tasks. The MRF estimates the parameters of a linear model for combining the view-based distinction maps. The experiments using several publicly accessible datasets show that the distinctive regions detected by our method are not just significantly different from those detected by methods based on handcrafted features, but more consistent with human perception. We also compare it with other perceptual measures and quantitatively evaluate its performance in the context of two applications. Furthermore, due to the view-based nature of our method, we are able to easily extend mesh distinction to 3D scenes containing multiple objects.

AB - An importance measure of 3D objects inspired by human perception has a range of applications since people want computers to behave like humans in many tasks. This paper revisits a well-defined measure, distinction of 3D surface mesh, which indicates how important a region of a mesh is with respect to classification. We develop a method to compute it based on a classification network and a Markov Random Field (MRF). The classification network learns view-based distinction by handling multiple views of a 3D object. Using a classification network has an advantage of avoiding the training data problem which has become a major obstacle of applying deep learning to 3D object understanding tasks. The MRF estimates the parameters of a linear model for combining the view-based distinction maps. The experiments using several publicly accessible datasets show that the distinctive regions detected by our method are not just significantly different from those detected by methods based on handcrafted features, but more consistent with human perception. We also compare it with other perceptual measures and quantitatively evaluate its performance in the context of two applications. Furthermore, due to the view-based nature of our method, we are able to easily extend mesh distinction to 3D scenes containing multiple objects.

U2 - 10.1109/TVCG.2018.2885750

DO - 10.1109/TVCG.2018.2885750

M3 - Article

JO - IEEE Transactions on Visualization and Computer Graphics

JF - IEEE Transactions on Visualization and Computer Graphics

SN - 1077-2626

ER -