TELKOMNIKA Telecommunication, Computing, Electronics and Control
An efficient method for stamps recognition using Haar wavelet sub-bands

TELKOMNIKA Telecommunication, Computing, Electronics and Control
An efficient method for stamps recognition using Haar wavelet sub-bands

Energy
Haar wavelet
Low order moment
Recognition rate
Stamp recognition

The problem facing certain organizations such as insurance companies and
government institutions where a huge amount of documents is handled every
day, hence an automated stamp recognition system is required. The image of
the stamp may be on a different background, with different sizes, and suffers
from rotating in different directions, also, the appearance of soft areas
(patches) or small points as noise. Thus, the main objective of this paper is to
extract and recognize the color stamp image. This paper proposed a method to
recognize stamps, by using a technique named Haar wavelet sub-bands. The
devised method has four stages: 1) extracts the stamp image; 2) preprocessing
the image; 3) feature extraction; and 4) matching. This paper is implemented
using C sharp (Microsoft Visual Studio 2012) programming language. The
experiments conducted on a stamp dataset showed that the proposed method
has a great capability to recognize stamps when using Haar wavelet transform
with two sets of features (i.e., 100% recognition rate for energy features and
99.93% recognition rate for low order moment).

Maha A. Rajab, Loay E. George

http://journal.uad.ac.id/index.php/TELKOMNIKA

Nov 25, 2020

peri irawan

pdf

english

text