Off-Line Handwriting Recognition with Context-Dependent Fuzzy Rules
Ashutosh Malaviya, F. Ivancic, J. Balasubramaniam, Liliane Peters · 2020
The problem of off-line word recognition involves uncertainties at various levels. The fuzzy logic methodology is employed for handwriting analysis and recognition according to various uncertainties, from low-level preprocessing and segmentation tasks to word interpretation according to context. This chapter introduces a multi-level fuzzy word recognition methodology. In the learning process, isolated or segmented handwriting portions are identified and stored in a form of knowledge base. The chapter describes a methodology that divides the problem of handwriting recognition into various levels. It also introduces a basic paradigm to describe pattern parameters. The chapter focuses on the feature extraction level of the proposed multilevel methodology. Fuzzy rule-based systems for pattern recognition and image understanding use a linguistic approach to overcome the information uncertainties. The main objective of the fuzzy aggregation mechanism is to find an overall measure for certain fuzzy information from uncertain and imprecise information data.