Comparative Study of Moments Shape Descriptors and propose a new hybrid Descriptor technique

Abstract

The shape is a significant visible feature of an object in images. Finding images by using shape features has gained much consideration. For precise computerized classification of the type of object in the image, different features extraction approaches may be used. To identify and determine the preferences among them, three different techniques are studied extensively and determined how it affected by the change which may happen to the different shapes with the proposal of a hybrid method that contains the robust descriptors of each technique which improves the results and provide strong features for shape description that achieves high accuracy reaching more than 90%.