A Method based on Decision Trees and Neural Network for Recognition of Farsi Handwritten Digits
| Main Author: | Navid Samimi Behbahan |
|---|---|
| Format: | Article eJournal |
| Bahasa: | eng |
| Terbitan: |
, 2013
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/3521131 |
Daftar Isi:
- Farsi handwrite character recognition is a main topic in pattern recognition, machine learning, image processing, machine vision and data mining. Handwrite character recognition has many applications such as licenses plate recognition, document annotation by keywords, postal code recognition, bank check processing, entry score system in university. In handwrite recognition confront with some difficult such as different type, written with different pressure, using a thick or thin pen. In general there are three major stages in the character handwrite recognition problem: (1) preprocessing that proceed to normalization, noise removing and segmentation, (2) Feature extraction tries to replace the image with numerical feature vector in order to describe image as well. (3) Classification phase try to recognition of handwrite character with high accuracy by using extracted feature. In this research, we using of Structural features and some of character numeral feature to recognition handwrite digit. Percent of recognition of this method for handwrite digits achieved 98.18%