Award Date


Degree Type


Degree Name

Master of Science (MS)


Computer Science

First Committee Member

Evangelos Yfantis

Number of Pages



The main focus of this research is to automate medical form processing. An important step in this process is separating handwriting from printed text characters. We developed a filtering technique that extracts handwritten text from the printed text in the form. Once the handwritten text is segregated, each line of the segregated text is identified. The identification step is followed by character segmentation. Statistical analysis is performed on the gaps between the characters in each line. This results in a binormal curve clearly depicting two regions indicating if the gap represents the spacing between characters within a word or between two words. Furthermore, an algorithm is employed for number recognition. We use different feature extraction algorithms and generate a high dimension feature vector. The algorithm is trained by giving training samples; a rule is generated to classify an input. A rule database is created in order classify the characters given during testing phase. By this method, there is no need to correlate the observed number with the pre-stored characteristics of numbers, instead we test the given number whether it satisfies the appropriate rule.


Handwritten; Number; Recognition

Controlled Subject

Computer science

File Format


File Size

2201.6 KB

Degree Grantor

University of Nevada, Las Vegas




If you are the rightful copyright holder of this dissertation or thesis and wish to have the full text removed from Digital Scholarship@UNLV, please submit a request to and include clear identification of the work, preferably with URL.


IN COPYRIGHT. For more information about this rights statement, please visit