The Computational Facets of Normalization Strategies for the Recognition of a Language
Source: Kurukshetra University
Normalization of a character image is a crucial preprocessing stage in the development of robust text recognizers. It helps in the standardization of size, position as well as font variants of a same character shape so as to reduce the differences between the images of the same character class. Comparison between the ratio-based normalization and scaling based normalization is made to analyze their affects in the process of standardization with respect to shape degradation and deformity, they introduce in the original shape. The effect of each normalization method on a character image is evaluated. In fixed aspect ratio mapping the normalized image plane is mapped to a square.