Bayesian shape recognition using Principle Component Analysis and Modified Chain Codes

This paper represents a shape recognition method using PCA (Principal Component Analysis) and Bayesian probability with MCC (Modified Chain Code). MCC is a shape descriptor which is invariant in 2D shape’s scale, translation and rotation. Shape prior information is analyzed by PCA using shape database. We recognized shapes by Bayesian probability with shape prior information. In this paper we describe traditional chain code to describe object features and using its modification version, PCA and Bayesian probabilistic analysis and classification are represented.