Principal features for Indian currency recognition

Abstract

Currency recognition system is one of the fast growing research fields under image processing. This paper proposes a novel method for Indian currency recognition. Our proposed approach identifies denomination by extracting features like Center Numeral, Shape, RBI Seal, Latent Image and Micro Letter. Principal Component Analysis is used to reduce the dimensions and a similarity based classifier is constructed to predict test sample. Results are also validated by constructing models using classifier implemented using WEKA and testing with unseen samples not considered in feature extraction. Our study demonstrated that center numeral results in an accuracy of 100% with all family of currencies.

A Hybrid Fake Banknote Detection Model using OCR, Face Recognition and Hough Features

“…A wide range of algorithms can be applied to the input data and problems such as the build-up of noise and signal distortion during processing can be avoided. In [9], it is mentioned that digital image processing permits the use of much more intricate algorithms, and hence, can handle both complicated and simple tasks, and even implement the methods which would be impossible by analog means. They will automatically reject counterfeits.…”

“…Another constraint is that the illumination conditions over the image must be uniform. In [9], they lacked using cross validation method such as which can be used for currency recognition to make the system more reliable. They developed a good currency recognition system in [17], but their proposed technique was unable to distinguish genuine notes from counterfeits.…”

Currency Recognition and Calculation System using Machine Learning Techniques

“…There are two methods used to get a better result such as image smoothened and image adjusted. Vishu R et al [11], there are five steps that are recommended such as preparation of dataset, prepare train, set feature projection, features construct classification model and prediction of unseen samples [12]. This paper focuses mainly on features extraction and detection of different currency notes.…”

Efficient Banknote Recognition Based on Selection of Discriminative Regions with One-Dimensional Visible-Light Line Sensor

“…There have also been studies combining both of the above feature extraction approaches. The Indian currency recognition method proposed by Vishnu and Omman [ 18 ] selects five regions of interest (ROI): the textures of center numerals, shapes, Reserve Bank of India (RBI) seals, latent images, and micro letters on scanned banknote images. PCA is subsequently used for dimensionality reduction of the extracted features.…”