An Efficient Skewed Line Segmentation Technique for Cursive Script OCR

Algorithm 2

Skew correction algorithm.

	Input: Noiseless skewed image.
	Output: Skew-less image.
	//Begin
	Step 1. Convert input text image into a grayscale image.
	Step 2. Scan the document and extract ROIs. // ROI = area between text lines.
	Step 3. Find all pixels between text lines (other than text pixels).
	Step 4. Join these pixels to fit on a line (having the same slope as text).
	Step 5. Go for the center of each fitted line.
	Step 6. Find the angle between the fitted line and the horizontal line of the page.
	Step 7. Rotate area.
	//End