Optical Character Recognition and Text to Speech Generation System using Machine Learning

A branch of computer science and artificial intelligence (AI), the study of machine learning seeks to mimic human learning and enhance accuracy over time using data and algorithms. Neural networks are now mostly employed for tasks involving pattern recognition. Picture capture of the handwritten paper and saving it in image format is an easy technique to store the data. “Optical Character Recognition” is a technique for converting handwritten data into electronic format. The science of optical character recognition makes it possible to convert different kinds of documents or photos into data that can be analyzed, edited, and retrieved. Some of the processes involved are pre-processing, segmentation, feature extraction, and post-processing. OCR has been utilized by numerous researchers to recognize characters. This project links handwritten recognition and integration and text-to-speech technology. CNN is used in this system's design to recognize the characters in a test dataset. The four steps of the suggested method are text extraction, text localization, text extraction, and text to speech conversion. The primary goal of this research is to examine CNN's capacity for character identification in the image dataset as well as the recognition accuracy after training and testing. Using multiclass Support Vector Machines (SVM), OCR is used to translate text from an image into editable text, and a Text to Speech system outputs audio.