Text Extraction and Detection from Images using Machine Learning Techniques: A Research Review

Abstract

Machine Learning is subset of Artificial Intelligence and there is lots of research growth across the world. It has capability to learn by its own without seeking any help from human beings or without any explicit programming based on its previous experience and knowledge. Mainly it can make its own decisions or can predict in performing certain tasks based on the res pective input dataset and its training set. Machine Learning is applicable in various real-time applications in our daily life. Text detection and text extraction is one of the important applications which contain valuable information from images captured from various sources. Text with different variations differs in their size, orientation, alignment, style, low brightness or contrast in images with complex background. Many facing issues in reading due to various variations in text on images. So, text detection and extraction is very important and challenging now a days. The objective here is to help human beings with a different language from different parts of the world so that they can easily read and understand any language written. Researchers use various machine learning algorithms and tools to recognize handwritten text and text captured from images to convert them into digital format. Optical Character Recognition (OCR) is a Machine learning technique that helps us to detect and extract text data or information of a document and turn it into editable and searchable data. This research paper mainly focusses on various machine learning algorithms that are applicable for text extraction of handwritten documents, images and detect them into digital format also translate it according to the user's requirements.