Concordia University

https://www.concordia.ca/content/shared/en/events/encs/computer-science/2020/03/31/Doctoral-Seminar-Saeed-Khazaee.html

LATEST INFORMATION ABOUT COVID-19

READ MORE

Examinations, Thesis defences

Doctoral Seminar: Saeed Khazaee (online)

Date & time

Tuesday, March 31, 2020
11:45 a.m. – 1:45 p.m.

Speaker(s)

Saeed Khazaee

Cost

This event is free

Where

Online

Speaker: Saeed Khazaee

Supervisor:
Dr. C. Y. Suen

Supervisory Committee:
Drs. N. Bouguila, T. Popa, W. Shang

Title: Detection of Counterfeit Coins Based on 3D Height-Map Image Analysis

Date: Tuesday, March 31, 2020

Time: 11:45 am

Place: Via Zoom teleconference

ABSTRACT

The main goal of this research is to propose a comprehensive method for analyzing 3D height-map images to detect counterfeit coins. We use 3D height-map images, since detecting a counterfeit coin using 2D image processing is nearly impossible in some cases, especially when the coin is damaged, corroded or worn out. In this research, instead of conventional 2-D methods for counterfeit coin detection, we applied a 3-D approach to model and analyze a large set of different Danish coins. One of the most important advantages of 3-D approaches is to extract features that cannot be resulted in 2-D images. As we encountered a lot of unexpected degradation and shadowing on shiny coin images, we faced wrong values of height or depth. To solve this problem, we, first, restored the degraded images without losing height information for which we provided two different proposed methods in our preliminary works. Secondly, we ignore the degradation problem for which in this research, we propose a new method to analyze the precipice borders that are not affected by the degradation problem. To do this, we propose a 3D approach to detect and analyze the precipice borders from the coin surface and extract significant features. In order to extract the features, we also propose Binned Borders in Spherical Coordinates (BBSC) to analyze different parts of precipice borders at different polar and azimuthal angles. Here, the system is trained and tested with four types of Danish and two types of Chinese coins. We also take advantage of stack generalization to classify the coins and add the reject option to increase the reliability of the system. The results illustrate that the proposed method outperforms other counterfeit coin detectors. The accuracy obtained by testing Danish 1990, 1991, 1996, and 2008 datasets are 98.6%, 98.0%, 99.8%, and 99.9% respectively. In addition, results for half Yuan Chinese 1942 and one Yuan Chinese 1997 were 95.5% and 92.2% respectively.

Back to top Back to top

© Concordia University