This paper is published in Volume-6, Issue-5, 2020
Computer Science Engineering
Lovepreet, Sarabjeet Kaur
Adesh Institute of Engineering and Technology, Faridkot, Punjab, India
Pub. Date
28 October, 2020
Paper ID
Skew Detection, Hough Transformation, Scan Line Algorithm, Document Skew Detection, Image Processing


Lovepreet, Sarabjeet Kaur. Skew recognition, International Journal of Advance Research, Ideas and Innovations in Technology,

Lovepreet, Sarabjeet Kaur (2020). Skew recognition. International Journal of Advance Research, Ideas and Innovations in Technology, 6(5)

Lovepreet, Sarabjeet Kaur. "Skew recognition." International Journal of Advance Research, Ideas and Innovations in Technology 6.5 (2020).


Skew detection has been an important part of the document recognition system. If the document includes the skew it will be difficult for reader to read the document. In the proposed work, new techniques that detecting the angle of skewed document. The main advantage is that the method works on all types of scripts. These techniques successfully handle the skew detection and correction of the printed and handwritten document. The detection and correction of document skew is one of the most important document image analysis steps. Page layout analysis and pre-processing operations used for character recognition depend on an upright image or, at least, knowledge of the angle of skew. A few methodologies have been proposed as choices for skew point identification of document images. Profile project technique is a popular method for skew detection. It is capable of locating fragmented lines in a binary image. Therefore given a group of black pixels, one can find the imaginary line or lines that go through the maximum number of these pixels. Existing techniques doesn’t work for images in which there is unequal space patterns in between the lines. Existing techniques cannot detect the skew in multiscript documents. Existing system is not capable of detecting correcting the upward and downward skew together in a single text document. Existing system Rotate the whole document at once to correct the skew from the given text document. A lot of work is required to be done in context of skew detection and correction for multiscript text images. These problems are required to be solved to make use of the system in the real world applications. The proposed system for skew detection and correction is implemented with the help of scan line algorithm along with hough transformation algorithm. Proposed system works on the document in line by line basis i.e. detect and correct the skew of every line independently. It can work on the documents having multiple scripts on it. It can detect and correct the skew from the documents even if there is an unequal space in between the two adjacent lines. It will rotate the document on line by line basis to correct the skew from the given document. Detects and correct the upward and downward skew together in the single text document. The major problem with the existing systems was that they just rotate the entire document in the reverse direction of the angle detected to remove the skew but the proposed system process the document on the line by line basis using scan line algorithm and skew angle is detected with the help of hough transformation. The performance of the proposed system is calculated on various images having different types of skew in them. The overall accuracy of the proposed system is 98% which is far better than that of existing system. In future the proposed system can be extended for coloured images and other types of documents. Proposed system can also be extended to made it for word level skew detection and correction to get more accuracy.
Paper PDF