A nonparametric approach to 3D shape analysis from digital camera images - I

JOURNAL OF MULTIVARIATE ANALYSIS

卷:101

Article

Patrangenaru, V.¹  Sugathadasa, S.² 

[1] Florida State Univ, Dept Stat, Tallahassee, FL 32306 USA

[2] Texas Tech Univ, Lubbock, TX 79409 USA

关键词: Pinhole camera images;    High level image analysis;    3D reconstruction;    Projective shape;    Extrinsic means;    Asymptotic distributions on manifolds;    Nonparametric bootstrap;    Confidence regions;   

DOI  :  10.1016/j.jmva.2009.02.010

来源: Elsevier

【 摘 要 】

This article for the first time. develops a nonparametric methodology for the analysis of projective shapes of configurations of landmarks on real 3D objects from their regular camera Pictures. A fundamental result in computer vision, emulating the principle of human vision in space, claims that, generically. a finite 3D configuration of points can be retrieved from corresponding configurations in a pair of camera images, Lip to a projective transformation. Consequently, the projective shape of a 3D configuration call be retrieved from two of its planar views, and a projective shape analysis can be pursued from a sample of images. Projective shapes are here regarded as points on projective shape manifolds. Using large sample and nonparametric bootstrap methodology for extrinsic means on manifolds, one gives confidence regions and tests for the mean projective shape of a 3D Configuration from its 2D camera images. Two examples are given: all example of testing for accuracy of a simple manufactured object using mean projective shape analysis, and a face identification example. Both examples are data driven based oil landmark registration in digital images. (C) 2009 Elsevier Inc. All rights reserved.

【 授权许可】

Free