Paper Title
Leech Step Path Finding Algorithm
Abstract
Sinhala characters inherit a variety of shapes especially different types of curves by nature. Therefore Sinhala
character recognition cannot reach the maximum accuracy with the existing techniques and algorithms used in Sinhala
character feature extraction within current intelligent character recognition (ICR), optical character recognition (OCR) and
optical music recognition (OMR) engines. Existing OCR engines even with their adaptive classifier have not been able to
train for Sinhala characters as it gets maximum accuracy with the output in the process of character recognition. In order to
solve this problem, we propose a novel algorithm, named as Leech Step Path Finding Algorithm (LSPFA) which can be
mainly used to extract/ recognize character features (of segmented/ isolated characters) in the adaptive classifier of ICR,
OCR and OMR engines. Even though our study principally targets Sinhala characters, LSPFA can be further used for any
other set of characters/ symbols of the alphabet of any other language.
Keywords - Character Feature Extraction; Sinhala Characters; Intelligent Character Recognition; ICR; Optical Character
Recognition; OCR; Optical Music Recognition; OMR; Adaptive Classifier