Recognition English Letters Using Invariant Moment in Extracting Properties of Images
AL-Rafidain Journal of Computer Sciences and Mathematics,
2005, Volume 2, Issue 1, Pages 69-82
AbstractIn this research, a study of printed English characters properties for Optical Character Recognition is done using the invariant moment method to evaluate the character properties due to it’s resistance against the rotation, scaling and shifting work. In this, also preprocessing the text and recognized it.The image text is segment often two dissuasion methods are used to measure the distance between the moments of the new characters and the stored moments, this method is the Euclidean distance which gives high recognition factor.Due to the property of independence of the English characters, the evaluation of the seven moment of the printed character is so easy, so we evaluate the moments of English character in a different position and different size for each character by rotating the character from 45° to 360°. As a result, the invariant moment method has high ability for the recognition use, if its adopted in English Optical Character Recognition the recognition rate equals 85%.
- Article View: 44
- PDF Download: 65