Optical and Handwritten Character Recognition System Using Segmentation and Neural Network

摘要

Optical character recognition has been a hot topic for decades and has been continuously being developed and improved. The following paper describes how English alphabets can be distinguished from one another based on process of segmentation and how density of marked contours helps neural networks to analyze each data. Based on experiments conducted on four skeletalized English alphabets, the following article highlights a brief description of how errors and anomalies occur during the process of handwritten character recognition.