Handwritten Cursive Character Recognition to Recognize Medicine Name on Medical Prescription

Abstract

" This thesis aims to represent the work related to English cursive character recognition in order to recognize medical prescriptions and automate the medicine availability checking process of a pharmacy. Cursive character recognition is really challenging because the characters are connected to each other. The proposed project is specifically for the medical domain to recognize medicine names from the prescription prescribed by the physician. The medicine names are recognized as an entire since there is no extra space between characters. The recognized text line segments are matched with a medicine name lexicon entries to acquire reliable segmentation and get valid medicine names. Finally, the recognized medicine names are matched with the available medicines lexicon in order to find the availability of the medicines. "