An Efficient Skewed Line Segmentation Technique for Cursive Script OCR

Algorithm 3

Text line segmentation algorithm.

	Input: Normalized de-skewed image
	Output: Segmented lines.
	//Begin.
	Step 1. //Preprocessing: image binarization (using adaptive threshold).
	Step 2. //De-skew the image (if needed).
	Step 3. //Scan Image row by row.
	Identify the intensity for each pixel (0 or 1).
	Step 4. //Calculate the standard deviation of the image (use as minimum black pixels in a text row).
	Step 5. If(Black_Pixels > Std)
	Black_Row = row
	Step 6. Else
	Space_Row = row
	Step 7. For (Start from 1st_Black_Row: till Last_space_Row)
	Step 8. If (Height_Row > Min_ Height_Row)
	//Consider these consecutive text rows as text line until any white_row occurs.
	Step 9. If (Space_Row occur)
	Step 10. If (Space_Row > Min_Height_Row)
	Break text_line and go to Step 11.
	Step 11. else
	Go to Step 7
	Step 12. else
	Search for next black_text_row
	Step 13. Else
	Go to Step 10
	//End