Impact of data augmentation on handwriting characters recognition
No Thumbnail Available
Date
2023
Journal Title
Journal ISSN
Volume Title
Publisher
University of Oum El Bouaghi
Abstract
Handwriting recognition involves converting handwritten text into digital form. It plays a pivotal role in various applications. Although recent advancements in optical character recognition have been significant, challenges continue to persist. Fortunately, the emergence of deep neural networks presents powerful solutions to address these obstacles. Motivated by the need for improved recognition accuracy in real-world scenarios, we aim to investigate the impact of data augmentation techniques on the performance of deep learning models for handwriting recognition tasks. The ensuing chapters of this dissertation delve into the fundamental principles of handwriting recognition, deep learning methodologies tailored for handwriting recognition, data augmentation techniques, and present comprehensive experimental analysis and results.
Employing a Convolutional Neural Network (CNN) model trained on two distinct handwriting datasets, this study conducts a comprehensive experimental analysis. The results underscore the efficacy of data augmentation in enhancing recognition accuracy, highlighting its potential to bolster the performance of deep learning models in handwriting recognition tasks.