CNN Implementation for Language Translation via Handwriting Image Recognition

Abstract

Abstract

This research explores the application of Convolutional Neural Networks (CNN) in handwriting image recognition for language translation purposes. With the increasing need to digitize documents, handwritten character recognition (HTR) systems have become very important. In this study, CNNs are used to convert handwritten images into digitized text with a high degree of accuracy, overcoming the challenges posed by variations in writing styles. The methods applied include data collection, image pre- processing, CNN implementation, and translation using Natural Language Processing (NLP). The results showed that the developed CNN model achieved a training accuracy of 94.73% and a testing accuracy of 90.53%. These findings show significant potential in improving the efficiency and accuracy of the process of translating handwriting into digital text. This research also notes recent advances in handwritten text recognition, including real-time character recognition and language-specific applications, demonstrating the relevance and wide application of this technology.