Please use this identifier to cite or link to this item:
|Title:||Automatic billing counterfeit detection for SGD money||Authors:||Arun Ramchandani.||Keywords:||DRNTU::Engineering::Electrical and electronic engineering::Electronic systems||Issue Date:||2010||Abstract:||Production of counterfeit notes is a crime carried out by professional criminals who acquire skills from current technology to produce flawless notes 100% accurate to the genuine bills. Cases have been highlighted as retailers and consumers have been duped due to counterfeit note scams. One of the biggest reasons why most counterfeit notes have been passed off as genuine ones is simply due to the cashier‘s failure to carry out basic checks especially during busy periods of large sales. This report will examine the feasibility of using the Automated Billing Counterfeit Detector for SGD money. The validation of the proposed design have been implemented based on the SIFT method. The discussed method would relate the techniques for extracting unique features that are invariant from the authentic Singapore Dollar notes that are used to have reliable matching performance between two counterfeit notes. Some features incorporate the invariance of scale and rotation, and provide robust matching over a long range. With a highly unique feature, meaning that any one feature can be matched perfectly when scanned against a large database of images and this would have a high probability. With object recognition also reflected in this paper, it can be viewed that recognition allows individual features that match to a large feature database. Currently, existing methods have known objects using an algorithm that have fast near neighbour. Following that, we have the Hough transform technique to identify single object extracted from clusters, and lastly having to verify through the least-squares solution with parameters that are consistent. This recognition approach can identify objects in a more robust manner among a large space of clutter while achieving a performance that is near real time. The design of the counterfeit system requires knowledge on image processing, detailed study of existing counterfeit detections using the different detection methods and basic understanding the different features of the Singapore denominations. The tests results show the detection of features using SIFT which incorporates image processing using MATLAB. This program gives its user the advantage of using a user-friendly detection system. It is applicable to both the paper portraits (1999 – 2003) and the current polymer based (2004 - present) denomination notes.||URI:||http://hdl.handle.net/10356/39414||Rights:||Nanyang Technological University||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||EEE Student Reports (FYP/IA/PA/PI)|
Page view(s) 20345
checked on Oct 23, 2020
checked on Oct 23, 2020
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.