Organic Shapes Classification by Similarity to Basic Geometric Shapes

This paper proposes and describes a novel technique for organic shapes classification by similarity to basic geometric shapes. The amber data used in experiments are collected by amber art craft industry experts and the presented investigations were care out in order to develop a classifier for online amber sorting application. The centroid distance function was selected for shape representation as it preserves the order of landmark points. The k-medoids and k-means clustering algorithms were compared by generating clusters of similar shapes for labeling to one of geometric shapes: circle, ellipse, oval, triangle, rectangle, rhombus, trapezium, and trapezoid. Clusters labeled by an expert to same categories were merged. Using labeled samples the decision tree classifier was trained. The training of classifier was made by acquiring all possible orientations of centroid distance function for each image in training set and then feeding them to decision tree. In the classification step all the shifted and flipped centroid distance function variations of the testing sample are voting for the class using the decision tree. Experimental results have shown that the proposed technique is effective in organic shapes classification to selected geometric shapes even if there is high ambiguity between organic shapes. Keywords-Expert systems, image classification, image matching, pattern clustering.