Automatic Incorporation of Corrections into Reviewed Documents

Abstract

In this paper we propose an image processing based solution to automate the task of incorporating suggested corrections into reviewed documents. The proposed method extracts the symbols put in by a reviewer on a hard copy and then recognizes the symbols to understand their meanings. The recognition is achieved by the use of centroidal profile features. A software package is developed to incorporate the corrections into the soft copy by identifying the positions of symbols (in the hard copy). Several experiments have been conducted to demonstrate the feasibility of the proposed methodology. Experimental results have demonstrated that the proposed method can withstand minor deformation and orientation and in addition, it is capable of handling joined symbols.