Mahidol University Logo
Faculty of ICT, Mahidol University
 

Admissions

Printable Version

 

IMAGE ENHANCEMENT FOR TAX FORM IMAGES

 

TITLE IMAGE ENHANCEMENT FOR TAX FORM IMAGES
AUTHOR PENTHIP CHAROENRIT
DEGREE MASTER OF SCIENCE PROGRAMME IN COMPUTER SCIENCE
FACULTY FACULTY OF SCIENCE
ADVISOR SUKANYA PHONGSUPHAP
CO-ADVISOR CHOMTIP PORNPANOMCHAI
 
ABSTRACT
This paper investigates methods of solving quality problems for tax form images. We considered five problems. The first problem is skew images; we proposed a method to rotate the skew tax form images into straight alignment by using a reference area of dilation text lines, and then we calculated the skew angle of image using that reference area. The second problem is salt-and-pepper noise images; we used median filter and kFill algorithm to remove salt-and-pepper noises. The third problem is marginal noise images; we proposed a method to remove that noise by using edge density to find a position of marginal noise, and then the marginal noise was deleted. The fourth problem is shadow backgrounds and low foreground contrast; we proposed a method to find the best parameters and applied a transformation equation for calculating a new gray level of pixel values. The last problem is overlap text/background on headline images; we proposed a method to identify the page of input tax form images by matching the control points of template and control points of input tax form images, using a correlation coefficient, and then replacing headlines of input tax form images with headlines from template. Experiments were performed on 20 tax form images for each method. Results are as follows. Skew correction has an average error 0.081 degrees. Salt-and-Pepper noises removal, marginal noise removal, shadow background removal and foreground contrast enhancement are satisfied by users at an average of 4.23, 4, and 4.5 of 5 points, respectively. For the overlap text/background on headline images correction, the page was correctly identified at an average of 95%, and the users were satisfied on average 4.57 out of 5 points. The results show that these methods have the potential to be applied to develop a system for solving the quality problems of tax form images.
KEYWORD TAX FORM IMAGE / IMAGE ENHANCEMENT/ SKEW IMAGE /MARGINAL NOISE/SALT-AND-PEPPER NOISE

 

Go to Top

 

ICT Building, Mahidol University, 999 Phuttamonthon 4 Road, Salaya, Nakhonpathom 73170 Tel. +66 02 441-0909 Fax. +66 02 849-6099
Mahidol University Computing Center, The Faculty of ICT, Mahidol University , Rama 6 Road, Rajathevi, Bangkok 10400 Tel. +66 02 354-4333 Fax. +66 02 354-7333