Robustness evaluation and implementation improvement of paper “Fingerprint” anti-counterfeiting system using flatbed scanner

• Main Author: Du, Qiu
• Other Authors: Qian Kemao
• Format: Final Year Project
• Language: English
• Published: 2017
• Subjects: DRNTU::Engineering::Computer science and engineering
• Online Access: http://hdl.handle.net/10356/70321
• Institution: Nanyang Technological University

Counterfeiting causes serious negative impact to the human society over the years. There are indeed some of anti-counterfeiting techniques used in practical. However, most of them have its own limitations. With the idea of using native surface texture of the paper, authentication of texture feature that paper introduces a new perspective of the anti-counterfeiting techniques. This novel idea is proposed by Clarkson et al. (2009). He proposed the idea of using commodity scanner to acquire texture information at each surface point of the paper from four orientations. And using normal vector information which is compute from that texture information for feature vector computation. This Final Year Project (FYP) is inspired by the framework proposed by Clarkson et al in 2009 and continued from previous student Li Zixuan who has done feasible study and a simple implementation at 2016. In addition to the great work of Li Zixuan who involves provides theory and simple prototype, the purpose of this FYP is to enhance the implementations and consummate the prototype to achieve goal of economic, accurate, robust, secure and widely applicable without modifying the original paper or adding additional artificial identifier/labels. In this FYP, some features that improve usability of the prototype has been implemented and some critical bugs has been corrected. For feature improvement, firstly, a simple yet much more user-friendly GUI has been created rather than just simple GUI of previous version. Secondly, a file protocol for image import has been added so that four different orientation images can be imported into application at once. Thirdly, corner detection and corner prediction feature has been implemented to aid users to easily select interested area of the paper. Last, cross resolution validation was evaluated on different paper. This paper “Fingerprint” anti-counterfeiting prototype has been partially implemented according to framework Clarkson et al. (2009) proposed. The core features have been successfully developed. Nevertheless, error correction bits and hash function are not implemented yet based on the framework. For the further improvements, possible future works are also advised in this report as potentials to achieve advance and wide applications.