A Comparative Analysis of Shape Based and Zernike Moment Feature Extraction Techniques for Fasteners Recognition Using Neural Network

Abstract

Abstract-This paper presents a Comparative Analysis of Shape-based and Zernike moment Feature Extraction Techniques for Fastener Recognition. There a nine features extracted using shape-based technique and 64 moments used in Zernike feature extraction technique. For Zernike moment technique, the 64 moments are divided into 3 groups. The first group is the lower order moments, the second group is the higher order moments and the third group is the combination between the lower order group and the higher order group. The processes taken in the recognition are image acquisition, pre-processing, segmentation, feature extraction, and classification. The segmentation process is carried out by using adaptive filter and the classification process employed artificial neural network. The final result from this experiment is that shape-based technique has a better classification result of about 84.93% correct recognition compared to Zernike moment technique which is about 51.53% (combined group).