Old and Worn Banknote Detection using Sparse Representation and Neural Networks

Seyed Aliakbar Mousavi; Majid Meghdadi; Zahra Hanifeloo; Putra Sumari   and Muhammad Rafie Mohd Arshad

doi:10.17485/ijst/2015/v8i10/53961

Article

Old and Worn Banknote Detection using Sparse Representation and Neural Networks

VIEWS 805
PDF 195

Abstract
Full-Text HTML
Full-Text PDF
How to Cite

Indian Journal of Science and Technology

DOI: 10.17485/ijst/2015/v8i10/53961

Year: 2015, Volume: 8, Issue: 10, Pages: 913-918

Original Article

Old and Worn Banknote Detection using Sparse Representation and Neural Networks

Seyed Aliakbar Mousavi ^1* , Majid Meghdadi ² , Zahra Hanifeloo² , Putra Sumari ¹ and Muhammad Rafie Mohd Arshad¹

¹School of Computer Science, Universiti Sains Malaysia (USM), Malaysia; [email protected]

²Computer Department, Faculty of Engineering, University of Zanjan, Iran

This work is licensed under a Creative Commons Attribution 4.0 International License.

Abstract

This paper provides a fast method to recognize a variety of Persian banknotes at different scales. In this technique, the PCA, LDA and sparse representation methods are utilized at feature extraction stage and follows with MLP neural networks, LVQ and SOM in classification. Finally, the application of sparse matrix representation method and combination of both SOM and LVQ neural networks would lead to the best efficiency with precision of 91.15% in recognition of Persian banknotes particularly the worn ones.

Keywords: Banknote Recognition, LVQ, Neural Network, PCA, Sparse Representation