Classification of two-dimensional shapes based on moment invariants

An experimental analysis of two-dimensional (2D) shape classification method based on moment invariants is presented. Various types of translation, scale and rotation invariants are used to construct feature vectors for classification. The performance is evaluated using five different objects picked up from real scenes with a TV camera. Silhouettes and contours are extracted from nonoccluded 2D objects rotated, scaled and translated in 3D space. The proposed feature extraction method are implemented and systematically tested using several parametric and nonparametric classifiers. The obtained results clearly justify the use of the 2D shape feature extraction method based on the moment invariants.