Edge detection using particle swarm optimization technique

Handwriting recognition has been a major topic of research since past many years. There have been several approaches to handwriting recognition but the recognition of handwritten characters using deep learning has been a hot topic of research in the past five years. This paper proposes a method of converting handwritten text into speech at real time using the concept of deep neural networks. Moreover the pre-processing is improved by using an enhanced edge detection method for thinning the boundaries of the segmented characters. The recognized text is converted into speech. This system provides the benefit of reducing the hectic task of manually going through the handwritten documents in places such as banks and post offices. It can have various applications such as the audio based checking of exam sheets of students in schools and colleges preventing the task of recognizing the handwritten characters, would be beneficial for the blind, for the people who are illiterate and do not know reading but know a specific language. It proposes to reduce the human effort in understanding one's handwriting and easily they can listen to the speech converted. It can be beneficial for the students as well as they can simple listen to their notes and learn better.