A Three-Phase Noise Removal Approach to Achieve Accuracy in Line Segmentation of Odia Text
Aradhana Kar, Sateesh Kumar Pradhan · 2021
This paper focuses on segmenting lines of a scanned document in a particular Odia font or a document typed in a particular Odia font which is converted to an image. When a document is scanned in a scanner or converted to image then, the resultant image consists of some noise along with the text. The approach described in this paper deals with segmenting the lines of the Odia text in an image by removing the noise. This approach consists of three modules: Preprocessing, RemoveNoise and LineSegmenting. A scanned document or a document of Odia text converted to image is given as input to Preprocessing which is responsible for converting the input image to gray scale image and then to binary image. The scanned document that is tilted has not been considered. The output of Preprocessing module is given as input to the RemoveNoise module which is responsible for removing the white spaces that surrounds the text and the noise present in the whole text. Then the LineSegmenting module takes the output of RemoveNoise module and produces segmented lines of the text.