Gaussian Process Deep Belief Networks: A Smooth Generative Model of Shape with Uncertainty Propagation

Alessandro Di Martino, Erik Bodin, Carl Henrik Ek, Neill Campbell

Research output: Contribution to conferencePaper

Abstract

The shape of an object is an important characteristic for many vision problems such as segmentation, detection and tracking. Being independent of appearance, it is possible to generalize to a large range of objects from only small amounts of data. However, shapes represented as silhouette images are challenging to model due to complicated likelihood functions leading to intractable posteriors. In this paper we present a generative model of shapes which provides a low dimensional latent encoding which importantly resides on a smooth manifold with respect to the silhouette images. The proposed model propagates uncertainty in a principled manner allowing it to learn from small amounts of data and providing predictions with associated uncertainty. We provide experiments that show how our proposed model provides favorable quantitative results compared with the state-of-the-art while simultaneously providing a representation that resides on a low-dimensional interpretable manifold.

Conference

ConferenceACCV2018 (Asian Conference on Computer Vision).
CountryAustralia
CityPerth, WA
Period2/12/186/12/18

Keywords

  • Shape Models
  • Gaussian Processes
  • Deep Belief Networks
  • Unsupervised Learning

Cite this

Di Martino, A., Bodin, E., Ek, C. H., & Campbell, N. (2018). Gaussian Process Deep Belief Networks: A Smooth Generative Model of Shape with Uncertainty Propagation. Paper presented at ACCV2018 (Asian Conference on Computer Vision)., Perth, WA, Australia.

Gaussian Process Deep Belief Networks: A Smooth Generative Model of Shape with Uncertainty Propagation. / Di Martino, Alessandro; Bodin, Erik; Ek, Carl Henrik; Campbell, Neill.

2018. Paper presented at ACCV2018 (Asian Conference on Computer Vision)., Perth, WA, Australia.

Research output: Contribution to conferencePaper

Di Martino, A, Bodin, E, Ek, CH & Campbell, N 2018, 'Gaussian Process Deep Belief Networks: A Smooth Generative Model of Shape with Uncertainty Propagation' Paper presented at ACCV2018 (Asian Conference on Computer Vision)., Perth, WA, Australia, 2/12/18 - 6/12/18, .
Di Martino A, Bodin E, Ek CH, Campbell N. Gaussian Process Deep Belief Networks: A Smooth Generative Model of Shape with Uncertainty Propagation. 2018. Paper presented at ACCV2018 (Asian Conference on Computer Vision)., Perth, WA, Australia.
Di Martino, Alessandro ; Bodin, Erik ; Ek, Carl Henrik ; Campbell, Neill. / Gaussian Process Deep Belief Networks: A Smooth Generative Model of Shape with Uncertainty Propagation. Paper presented at ACCV2018 (Asian Conference on Computer Vision)., Perth, WA, Australia.
@conference{b1beb1889699422a9d90fdd854644fb8,
title = "Gaussian Process Deep Belief Networks: A Smooth Generative Model of Shape with Uncertainty Propagation",
abstract = "The shape of an object is an important characteristic for many vision problems such as segmentation, detection and tracking. Being independent of appearance, it is possible to generalize to a large range of objects from only small amounts of data. However, shapes represented as silhouette images are challenging to model due to complicated likelihood functions leading to intractable posteriors. In this paper we present a generative model of shapes which provides a low dimensional latent encoding which importantly resides on a smooth manifold with respect to the silhouette images. The proposed model propagates uncertainty in a principled manner allowing it to learn from small amounts of data and providing predictions with associated uncertainty. We provide experiments that show how our proposed model provides favorable quantitative results compared with the state-of-the-art while simultaneously providing a representation that resides on a low-dimensional interpretable manifold.",
keywords = "Shape Models, Gaussian Processes, Deep Belief Networks, Unsupervised Learning",
author = "{Di Martino}, Alessandro and Erik Bodin and Ek, {Carl Henrik} and Neill Campbell",
year = "2018",
language = "English",
note = "ACCV2018 (Asian Conference on Computer Vision). ; Conference date: 02-12-2018 Through 06-12-2018",

}

TY - CONF

T1 - Gaussian Process Deep Belief Networks: A Smooth Generative Model of Shape with Uncertainty Propagation

AU - Di Martino, Alessandro

AU - Bodin, Erik

AU - Ek, Carl Henrik

AU - Campbell, Neill

PY - 2018

Y1 - 2018

N2 - The shape of an object is an important characteristic for many vision problems such as segmentation, detection and tracking. Being independent of appearance, it is possible to generalize to a large range of objects from only small amounts of data. However, shapes represented as silhouette images are challenging to model due to complicated likelihood functions leading to intractable posteriors. In this paper we present a generative model of shapes which provides a low dimensional latent encoding which importantly resides on a smooth manifold with respect to the silhouette images. The proposed model propagates uncertainty in a principled manner allowing it to learn from small amounts of data and providing predictions with associated uncertainty. We provide experiments that show how our proposed model provides favorable quantitative results compared with the state-of-the-art while simultaneously providing a representation that resides on a low-dimensional interpretable manifold.

AB - The shape of an object is an important characteristic for many vision problems such as segmentation, detection and tracking. Being independent of appearance, it is possible to generalize to a large range of objects from only small amounts of data. However, shapes represented as silhouette images are challenging to model due to complicated likelihood functions leading to intractable posteriors. In this paper we present a generative model of shapes which provides a low dimensional latent encoding which importantly resides on a smooth manifold with respect to the silhouette images. The proposed model propagates uncertainty in a principled manner allowing it to learn from small amounts of data and providing predictions with associated uncertainty. We provide experiments that show how our proposed model provides favorable quantitative results compared with the state-of-the-art while simultaneously providing a representation that resides on a low-dimensional interpretable manifold.

KW - Shape Models

KW - Gaussian Processes

KW - Deep Belief Networks

KW - Unsupervised Learning

M3 - Paper

ER -