Identification of Handwritten Simple Mathematical Equation Based on SVM and Projection Histogram
Sanjay S. Gharde1, Pallavi V. Baviskar2, K. P. Adhiya3
1Sanjay S. Gharde, is currently working as assistant professor in Computer Engineering Department, North Maharashtra University/ SSBT College of Engineering and Technology, Jalgaon India.
2Pallavi V. Baviskar is currently pursuing master’s degree in computer science engieeringin from North Maharashtra University, Jalgaon, India
3K. P Adhiya is currently Associate Professor in Computer Engineering Department of SSBT College of Engineering and Technology, Jalgaon, India.
Manuscript received on April 03, 2013. | Revised Manuscript received on April 28, 2013. | Manuscript published on May 05, 2013. | PP: 425-429 | Volume-3, Issue-2, May 2013. | Retrieval Number: B1579053213/2013©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Recognition of simple mathematical equation can applied on off-line handwritten samples. For smooth implementation database is prepaid with total 237 symbols which are collected from 28 different simple mathematical equations. The dataset 1and dataset 2 are training using most popular classifier named Support Vector Machine. In particular, this work tries to spotlight on evaluation of various methods used for feature extraction and recognition system. Moreover, some essential issues in simple mathematical handwritten equation recognition will be addressed in deepness. This paper discusses various steps of recognition process for simple mathematical handwritten equations. In that, pre-processing, segmentation, feature extraction, classification and recognition for handwritten mathematical symbol as well as for simple expression is described. Among the different phases applied in recognition system, features extraction and classification method may influence the overall accuracy and recognition rate of the system. Therefore, various techniques applied in this context are studied and comparative analysis is prepared. This evaluation study suggests projection histogram most suitable feature extraction technique and support vector machine is appropriate classification technique for implementation. Using projection profile and support vector machine two different dataset are recognized then 97.58% and 98.40% (as an average it resulted into 98.26%) recognition rate is achieved for simple handwritten mathematical equation.
Keywords: Classification, mathematical expression, projection histogram, support vector machine.