Abstract :

A number of applications emerging related to image to text conversion or image to speech conversion which helps us in various ways. This paper focussed on the conversion of Image-to-Speech in two different phases. The supporting tools used to implement this work are tesseract-OCR, machine learning based libraries, and speech recognizer using Python. For a given image the texted regions are partitioned using morphological operations. An annotated image to text and speech synthesis using cloud API is done. This work focused on the conversion of input image into text, and further the text is converted into English speech and then translated into speech of various Indian languages. The proposed system is useful for visually impaired persons who are not able to read or recognize the text given in many sign boards or images, and also in many more applications where we are in need of Image-to-Speech synthesis. It also focused on translating the English speech into a number of Indian languages such as Nepali, Urdu, Tamil, Malayalam, Bengali, Gujarati, Hindi, Kannada, and Marathi.