Continuous Generalized Procrustes analysis

作者：

Highlights：

• Continuous formulation of the generalized Procrustes analysis.

• CGPA avoids the need to generate 2D projections from all 3D rigid transformations.

• CGPA builds an efficient non-biased 2D shape model from 3D objects.

• CGPA uses the Haar measure to integrate over the space of 3D rotations.

• Experimental results building 2D shape models for different public datasets.

摘要

Highlights•Continuous formulation of the generalized Procrustes analysis.•CGPA avoids the need to generate 2D projections from all 3D rigid transformations.•CGPA builds an efficient non-biased 2D shape model from 3D objects.•CGPA uses the Haar measure to integrate over the space of 3D rotations.•Experimental results building 2D shape models for different public datasets.

论文关键词：Procrustes analysis,2D shape model,Continuous approach

论文评审过程：Received 23 July 2012, Revised 23 July 2013, Accepted 3 August 2013, Available online 13 August 2013.

论文官网地址：https://doi.org/10.1016/j.patcog.2013.08.006