Review on Currency Number Recognition

Reappraisal on Currency Number Recognition

Abstraction

Over the past old ages, a great technological progresss in colour printing, duplicating and scanning, forging jobs arrived. In the yesteryear, merely the printing house has the ability to do imitative paper currency, but today merely by utilizing a computing machine and optical maser pressman at house, it is possible to publish imitative bank notes. Therefore the issue of expeditiously separating forgery bills from echt via automatic machines has become more of import. Counterfeit notes are job of every state. Thus such a system is required, which is helpful in confirmation and acknowledgment of paper currency notes with fast velocity and less clip demand. These currencies will be verified by utilizing image processing techniques. This consists of image processing with feature extraction of paper currency. Image processing includes the nature of an image to better its ocular information for human reading. The consequence will be whether currency is echt or forgery.

General Footings

• Image Processing:

Digital image processing has become of import in many Fieldss of research, industrial and military applications. The processing on planar informations, or images, utilizing a digital computing machine or other digital hardware.

• Feature Extraction:

Feature extraction method is for bettering velocity and truth between two factors. Most normally used characteristic extraction method is image processing. It effects on design and public presentation of the system intensively.

Keywords

MATLAB Image Processing Toolbox, GUI ( Graphical User Interface )

1. Introduction

Feature extraction of images is the disputing work in digital image processing. The feature extraction of Indian currency notes involves the extraction of characteristics like consecutive Numberss, watermarking of currency. Feature extraction is that of pull outing the natural information from the given information. Probabilities of paper currencies with assorted states are likely rises progressively. This is a challenge for conventional paper currency acknowledgment systems. The acknowledgment of the consecutive Numberss of the Indian paper currency such as 100, 500 or 1000 can be detected utilizing assorted methods. The consecutive Numberss are used as identifiers that average IDs of bills.

2. CURRENCY RECOGNITION METHODS

2.1 A Reliable Method for Paper Currency RecognitionBy Junfang Guo, Yanyun Zhao, Anni Cai, IEEE Transactions, Proceedings of IC-NIDC2010,978-1-4244-6853-9/10. A Reliable Method for Paper Currency Recognition is based on LBP that means traditional local double star form ( LBP ) method, an improved LBP algorithm, besides called block-LBP algorithm, which is used for characteristic extraction. LBP tool is used for texture description. Advantages of this method have simpleness and high velocity.

2.2 Feature Extraction for Paper Currency RecognitionH. Hassanpour, A. Yaseri, G. Ardeshiri aˆ•Feature Extraction for Paper Currency Recognition, IEEE Transactions, 1-4244-0779-6/07,2007. In the techniques for paper currency acknowledgment, three features of paper currencies include size ; colour and texture are used in the acknowledgment. By utilizing image histogram, with the mention paper currency plenty of different colourss in a paper currency is computed and compared.

2.3 Feature Extraction for Bank Note Classification Using Wavelet Transform

Euisun Choi, Jongseok Lee and Joonhyun Yoon presented this paper in March, 2006 at IEEE International conference.In this paper probe to have extraction for bank note categorization by working the ripple transform. In the proposed method, high frequence coefficients taken from the ripple sphere and are examined to pull out characteristics. We foremost perform border sensing on measure images to ease the ripple characteristic extraction. The characteristic vectors is so conducted by thresholding and numeration of ripple coefficients. The proposed characteristic extraction method can be used to sorting any sort of bank note. However, in this paper scrutiny of Korean won measures of 1000, 5000 and 10000 won types. The textured parts of different measure images can be easy described by break uping the texture into several frequence sub-bands. In the proposed method, high frequence bomber sets are explored to pull out characteristics from transformed images.

2.4 Texture Based Recognition Techniques

Texture is a most utile characteristic for Currency acknowledgment. Textural characteristics related to human ocular perceptive are really utile for characteristic choice and texture analyser design. There are some set of texture characteristics that have been used often for image retrieval. Tamura characteristics ( saltiness, directivity, contrast ) , Tamura saltiness is defined as the norm of coarseness steps at each and every pel location inside a texture part. These characteristics can calculate straight from the full image without any similarity. In general the public presentations of this characteristic are non satisfactory. The saltiness information utilizing a histogram should be considered. The Gabor characteristic usage filters to pull out texture information at multiple graduated tables and orientations. As for texture characteristics, there is a comparing of the public presentation of Tamura characteristics, border histogram, MRSAR, Gabor texture characteristic, and pyramid-structured and tree-structured ripple transform characteristics. Harmonizing to author the experimental consequences indicated that MRSAR and Gabor characteristics perform other texture characteristics. However, to accomplish such good public presentation from MRSAR, the Mahalanobis distance based on an image-dependent Covariance matrix has to be used and it increases the size of characteristic and hunt complexness. The extraction of Gabor characteristic is much slower than other texture characteristics, which makes its usage in big databases. Generally Tamura characteristics are non every bit good as MRSAR, Gabor, TWT and PWT characteristics.

2.5 Placement Rule

In the yesteryear, there were some troubles in texture analysis due to miss of equal tools to qualify different graduated tables of texture efficaciously. There are some texture based techniques. The work done in this country was carried out by Tamura. Harmonizing to him, for ocular texture is hard. Its construction is attributed to the insistent forms in which elements are arranged harmonizing to a arrangement regulation. Hence it can be written as f= R ( vitamin E ) , Where R is denoting a arrangement regulation ( or relation ) and e is denoting an component. There is a set of characteristics utilizing this all input forms are measured and gives good distributed consequences. So it is required to hold both extremes defines for each characteristic. e.g. , harsh versus mulct for saltiness. Coarseness is a extremely of import factor in texture. In order to better the other characteristics, its consequences should be utilized.

2.6 Pattern Based Recognition Techniques

The Pattern acknowledgment is based on anterior cognition as a characteristic. This is the categorization of objects based on a set of images. These techniques are focused on Vector quantisation based histogram mold. Vector quantisation ( VQ ) is a method of trying a d-dimensional infinite where each point,tenJ, in a set of informations is replaced by one of the L paradigm points. The paradigm points are selected such that the amount of the distances ( deformation ) from each information point,tenJ, to its nearest paradigm point is minimized. The work in this country was completed out by Seth McNeillIn et Al. Author gives the method for acknowledgment of coins by pattern acknowledgment. This differentiates between the bald bird of Jove on the one-fourth, the torch of autonomy on the dime, Thomas Jefferson ‘s house on the Ni, and the Lincoln Memorial on the penny. First collects the information, during the informations aggregation phase assorted background colourss, including black, white, ruddy, and blue, were tested for segmentability. Adobe Photoshop was used to find the RGB values of the coin and its background. Then Segmentation was applied to these images. After the informations aggregation next is Coin Segmentation and Cropping. In this measure coins were segmented from their backgrounds by utilizing some alteration of Nechyba’s codification. Croping plan was implemented to turn up the borders of coin. After this Features were extracted from the coins by texture templets with each image, with border sensing templets. and The consequence of this method is 94 % accurate.

2.7 Color Based Recognition Technique

The Wei-Ying Maetal. in describes Color histogram ( CH ) method for an image. It is created by numbering the figure of pels of each colour. Histogram describes the colour distribution in an image. It is easy to calculate and is insensitive to little alterations in sing place ( VP ) . The calculation of colour histogram involves numbering the figure of pels of specified colour. Therefore in an image with declaration m*n, the clip complexness of calculating colour histogram is O ( manganese ) . It overcomes some of the jobs with colour histogram techniques such as high-dimensional characteristic vectors, spacial localisation, and indexing and distance calculation.

3. SYSTEM OVERVIEW

3.1 Flow of Image Processing

Fig 1. Flow of System

This system is designed by using image Processing tool chest and other related Matlab tool chest. The system is divided into some subdivision to back up the hereafter acknowledgment procedure.

4. Recognitions

A thesis work of such a great significance is non possible without the aid of several people, straight or indirectly. First and foremost I have huge felicity in showing my sincere thanks to my usher, Prof. Vishal Bhope for his valuable suggestions, co-operation and uninterrupted counsel. I am really much thankful to all my module members.

5. Reference

[ 1 ] Hanish Aggarwal and Padam Kumar, “Localization of Indian Currency Note in Color Images” , ICCCNT 2012. ( Unpublished ) .

[ 2 ] Wei-Ying Ma and HongJiang Zhang, “Benchmarking of Image Features for Content-based Retrieval” Hewlett-

Packard Laboratories, 1501 Page Mill Road, Palo Alto, CA 94304-1126.

[ 4 ] Hideyuki Tamura, Shunji Mori, and Takashi, “Textural Features Matching to Visual Perception” , Member IEEE.

[ 5 ] Seth McNeill, Joel Schipper, Taja Sellers, Michael C. Nechyba“Coin Recognition utilizing Vector Quantization and Histogram Modelling” Machine Intelligence Laboratory University of Florida Gainesville, FL 32611.

[ 6 ] Michael C. Nechyba, “Vector Quantization a confining Case of EM” , EEL6825: Pattern Recognition Class Material, Fall 2002.

[ 7 ] Jing Huang, S Ravi Kumar, Mandar Mitra, Wei-Jing Zhu, Ramin Zabi, “Image Indexing Using Color Correlograms” , Cornell University Ithaca, NY 14853.

[ 8 ] John R. Smith and Shih-Fu Chang, “Tools and Techniques for Color Image Retrieval” , Columbia University Department of Electrical Engineering and Centre for Telecommunications Research New York, N.Y. 10027.