“…There is a need to develop word level skew angle estimation algorithm as the skew may vary even in a single word due to handwriting variations. Most of the implemented algorithms works for printed documents and their performance for handwritten documents is very poor [39]. Range of the estimated skew angle is also limited.…”