An Experimental Exploration of Segmentation Techniques for Modi Script

The segregation of image is most significant phase to identify borderline information of an image effectively. As various degradation exists in an image such as distorted pixel value, blurring in an image, poor luminance etc that affects the visual representation of an image. Using segmentation techniques, we attempts to improve the content of an image and make it clear for representation. There exist various edges and clustering based segmentation techniques such as perwitt, roberts, canny, sobel and K-means clustering that assist in segregating distortion information from a Modi character image to great extent. The comparative analysis of these segmentation techniques based on some performance parameters is performed to segment Modi character components. As a result, Kmeans clustering technique shows more appropriate outcome for segregating Modi numerals efficiently.