A Hybrid Approach To Detect And Correct A Skew In Scanned Document Images Using Fast Fourier Transform And Nearest Neighbor Algorithm
Document image processing has become an growingly important technology in the automation of office
documentation tasks. When a document is scanned through automatic document scanners either mechanically or manually
for digitization, it often suffers from some degree of skew or tilt. This paper describes two algorithms to estimate the text
skew angle in a document image. The skew angle is obtained by looking for a peak in histogram of the gradient orientation
or Fourier analysis of the input grey-level image. Then the skew image is corrected by a rotation at such an angle which is
estimated by the algorithm. The method is not limited in the range of detectable skew angles and the achievable accuracy. To
measure the processed time and speed taken by skew detection algorithm, the Fast Fourier Transform technique is applied as
it is fast approach for finding the angle of skewed document. Nearest neighbor method is founded on connected components
in which the first nearest-neighbors of all connected components are and the histogram of the direction vectors for all
nearest-neighbors is acquired. Finally, in order to evaluate the performance of the proposed methodology we compare the
experimental results with those of well-known existing methods.
Keywords- Skew Detection, Scanned Documents, Fast Fourier approach, nearest neighbour approach, Skew Correction.