Multi-script Text Detection from Images: A Survey

Abstract

Textual information in images contains small but crucial information. Text detection is first necessary step to extract text from images. It has applications including intelligent car driving system, translation, information retrieval and many more. Therefore as a result, it is broadly studied. Detecting texts from images is a complex problem due to background, varied font size, style, orientation and transparency of text. Detecting multilingual text from images adds further complexity to computer vision problem that is primary to a collection of prominent applications. Since India is a multilingual country, multi script texts appears almost everywhere. The universal features for multi script environment are unknown due to different format, height and strokes of languages. Therefore detecting multi script text from images is an important yet unsolved problem. This paper provides a survey on available techniques to detect multi script text from images.