Related Experiment Video
Updated: Mar 22, 2026
06:37
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
Published on: December 15, 2023
5.6K
A hybrid recognition system for off-line handwritten characters
Gauri Katiyar1, Shabana Mehfuz2
1Department of Electrical Engineering, Jamia Millia Islamia, New Delhi, India ; ITS Engineering College, 46 Knowledge Park, Greater Noida, Uttar Pradesh 201308 India.
Springerplus
|April 12, 2016
Summary
This study introduces a novel approach for handwritten character recognition by combining multiple features and optimizing them with a Genetic Algorithm. This method enhances accuracy and reduces computation time for pattern recognition systems.
Area of Science:
- Computer Science
- Artificial Intelligence
- Machine Learning
Background:
- Pattern recognition involves crucial sub-processes like feature extraction and selection.
- Selecting optimal features is complex and vital for effective pattern recognition systems.
Purpose of the Study:
- To develop an efficient handwritten character recognition system.
- To improve accuracy and reduce computational time in pattern recognition.
Main Methods:
- Combined multiple features from seven different extraction approaches.
- Utilized a Genetic Algorithm for feature optimization.
- Employed a Multi-Layer Perceptron classifier for recognition.
Main Results:
- Achieved higher accuracy in handwritten character recognition.
- Demonstrated reduced computational time compared to traditional methods.
- Validated effectiveness using the CEDAR database for English alphabet recognition.
Conclusions:
- The proposed system effectively recognizes handwritten characters.
- Combining optimized features with a Multi-Layer Perceptron classifier is a viable approach.
- This method offers a promising solution for efficient pattern recognition.
