Local Slant Estimation and Correction for Handwritten English Words

Abstract

Handwritten words often have slant which need to be estimated and corrected before character segmentation and recognition. Several methods have been proposed for average slant estimation and correction. However, average slant estimation has the problem such that local slant will be overestimated or underestimated when the slant in a word varies from character to character. To solve the problem, this paper proposes three methods for local slant estimation, which are simple iterative method, high speed iterative method and 8-directional chain code method. The experimental results show that the proposed methods can estimate and correct local slant more accurately than the average slant correction.