Doctoral Seminar: Ali Al-Frajat

Speaker: Ali Al-Frajat

Supervisor: Dr. C. Y. Suen

Supervisory Committee:
Drs. L. Lam, C. Poullis, W-P. Zhu

Title: Scale and Rotation Invariant Coin Recognition and Authentication Based on Features of Segmented Characters

Date: Friday, March 31, 2017

Time: 10:15 am

Place: EV 3.309

ABSTRACT

Tremendous amount of coins are widely used in our daily life since ancient times. The demands of an efficient and reliable computer-aided system to classify those coins have escalated since. Coin recognition received enormous attention in recent years where two big workshop competitions took place in 2006 and 2007 organized by MUSCLE CIS-Benchmark. The goal of the competitions is to classify European coins from 12 countries after the introduction of Euro coins. Therefore, the focus of this research is to investigate coin features that uniquely identify coins. The proposed solution introduces a dynamic character recognition and segmentation method from coin images based on gradient features of each character. The extracted feature set is used to classify characters and recognize the coin. The proposed solution also considers the size of extracted features and applies a dimensionality reduction method to reduce the feature vector size and improve the classification results. A set of experiments was conducted on different datasets to demonstrate the effectiveness of the proposed solution.

On the other hand, a major issue with coins throughout the years is forgery. There are large numbers of spurious coins in circulation around the world. Governments are seeking reliable solutions to overcome the counterfeiting problem. In this research, we are aiming to develop a coin authentication system considering several security features on the coin’s surface along with their measures and characteristics.