Text Extraction and Language Translation for Health Care and Clinical Records Using Deep Learning

Abstract

Language conversion is a technique that has tremendous use worldwide. With the multitude of languages available, it is not feasible for an average person to learn languages he/she will not use. However, everyone faces situations where knowing a certain language would help in a certain city or even during work. This problem can be solved using the latest technologies in deep learning and computer vision. A system that can identify the text from images and convert them into a well-known language like English has tremendous use everywhere especially in the medical field. The medical field is one where communication and exchange of information is critical. In case of a medical emergency during a tourists visit, the inability to understand what the doctor is trying to say or what type of medication is being prescribed would lead only to more stress and panic. There are systems that have tried to solve this problem and there will be more in the future. One of the glaring issues faced by contemporary systems is the lack of image processing before getting the text of a signboard using computer vision algorithms. Blurred images or images where the characters of the text are too close to each other can hamper the system’s performance to a great extent. The proposed system provides a way by which the text present in an image can be extracted and displayed back to the user in a universal language like English. The system incorporates an optical character recognition tool called pytesseract to get the text from the image and it is converted into a different language using deep learning.