A Handwriting Recognition System That Outputs Editable Text And Audio

A digitalized version of the text is needed when preparing the audiobooks and presentations for review. Compared to a text file, understanding handwriting is quite challenging, so converting the handwritten text to digitalized text makes it easier for the readers to understand. The audio version of the text file is made accessible to blind people. It is essential to use digitalized text when performing Natural Language Processing and data analytics techniques. Handwriting recognition (HWR), also known as handwritten text recognition (HTR), is the ability of a computer to receive and interpret intelligible handwritten input from sources such as paper documents, text images, touch screens, and other devices. An image of a word, a line or a paragraph of text, or even a full document is analyzed, and the sequence of characters that compose the text is expected as output. The output obtained by this form is regarded as a static representation of handwriting that can be used within computer and text-processing applications. Now-a-days, Deep learning has proven its functionalities in a wide range of domains, including sounds, images, videos, and Natural Language Processing. Deep Learning techniques such as Convolutional Neural Networks (CNNs), Optical Character Recognition (OCR), and Long Short-Term Memory (LSTM) are playing a key role in handwriting recognition. For this experimentation, the IAM dataset is used for model training and testing. The performance of the model is typically measured by the “recognition rate” and “character error rate” that will be analyzed in this project.