應用隱藏馬可夫模型於偽鈔辨識之研究

The Application of Hidden Markov Models in Counterfeit Banknote Recognition

劉國書(Kuo-Shu Liu)

指導教授 ： 陳文輝

國立臺北科技大學/機電學院/自動化科技研究所/碩士(2012年)

摘要

由於科技進步，電腦、取像及列印設備精良且使用相當容易，導致偽鈔案件層出不窮，大量偽鈔流入市面，造成經濟損失與金融紊亂。而許多場合所依賴的自動辨識機器並非檢測鈔票內所有防偽項目，在面對製作精良的偽鈔，可能造成機器的誤判，因此發展高效能偽鈔辨識系統已是刻不容緩。本研究提出以機器視覺結合隱藏馬可夫模型，實際開發一套偽鈔辨識系統，利用紙張與浮水印的特性，使用背光源、影像定位與飽和度色彩空間轉換獲取難以仿造的鈔票特徵影像，並透過主成分分析法取得特徵影像之特徵向量以加速系統處理的時間，最後使用具有辨識時序能力的隱藏馬可夫模型作為系統的辨識核心。實驗結果顯示在有限的樣本下，皆能準確辨識真鈔與偽鈔，期望能使得現有防偽機制更加完善。

關鍵字

偽鈔辨識 ； 特徵擷取 ； 隱藏馬可夫模型

並列摘要

The improvement in image capturing technologies and the development of high-precision printing equipments makes it relatively easy to produce counterfeit notes. Large amount of counterfeit notes in the market causes economic loss and financial disorder. The automatic identification machines used in many cases are unable to detect all of the security features. Perfectly printed counterfeit notes result in the misjudgment of identification machines. So the development of high-performance recognition system is an urgent issue. This research proposed a machine vision combined with the hidden Markov model to develop an actual system that identifies counterfeit notes. With the characteristics of the paper and the watermark, utilizing backlight, image positioning, and saturation space conversion to acquire the image of the currency note characteristic that is difficult to forge, and also, through the method of principal component analysis to acquire feature vector of the characteristic image to accelerate the system processing time. Finally, utilize the hidden Markov model with sequence identification capability as the identification core of the system. The result show that under limited number of samples, the system can accurately identify genuine and counterfeit notes, hoping to enhance the current counterfeit identification mechanism.

並列關鍵字

Counterfeit Banknote Recognition ； Feature Extraction ； Hidden Markov Model

參考文獻

[4] C.Y. Yeh, W.P. Su, and S.J. Lee, “Employing multiple-kernel support vector machines for counterfeit banknote recognition,” Applied Soft Computing, vol.11, pp. 1439-1447, 2011.

[9] C.X. Liu, S.C. Ruan, G.M. Huang, Y.B. Jian, and L. Zhang , “Research on identification the counterfeit by recognizing the infrared image,” IEEE International Conference on Microwave and Millimeter Wave Technology, vol.4, pp. 2081-2084, 2008.

[12] H. Hassanpour, and P.M. Farahabadi, “Using hidden markov models for paper currency recognition,” Expert Systems with Applications, vol.36, pp. 10105–10111, 2009.

[13] F. Grijalva, J.C. Rodriguez, J. Larco, and L. Orozco, “Smartphone recognition of the U.S. banknotes denomination for visually impaired people,” IEEE Conference on ANDESCON, pp. 1-6, 2010.

[17] R.M. Haralick, K. Shanmugam, and I.H. Dinstein, “Textural features for image classification,” IEEE Transactions on Systems, Man, and Cybernetics, vol.3, pp. 610–621, 1973.