Enhancement of Sift algorithm to check authenticity of Indian Currency

摘要

The image analysis is a technique of image processing that helps to analyze various properties of the input image. In the recent times various application of the image analysis has come into existence. The current detection is the application of image analysis which is used to define that whether currency is fake or genuine. Various algorithms has been proposed based on SIFT algorithm to analyze external features of input currency. In the existing algorithm, only external features are analyzed that has accuracy of detection. In the proposed work, new technique has been proposed to analysis on to analyze internal features of the image for the detection. The internal features of the image are strip values of the currency. Simulation has been performed in MATLAB. Peak signal to noise ratio (PSNR), root mean square error (RMSE), fault detection rate (FDR) and accuracy values of the proposed and existing algorithm are compared to validate the proposed system. It is been found that accuracy of detection has increased up to 10 percent and fault detection rate is reduced up to 8 percent.