TEXT DETECTION AND LOCALIZATION IN IMAGES AND VIDEOS

Abstract

Extraction and recognition of text present in video and images has become a very popular research area in the last decade. Generally, text present in images or video frames is of different size, orientation, style, etc. with complex backgrounds, noise, low resolution and contrast. These factors make the automatic text extraction and recognition in images or video frames a challenging task. In general text present in images and videos can be classified in to scene text and overlay text. Scene text occurs naturally while overlay text is added artificially. The content of the scene or editor's intention can be well understood by Overlay text. Also overlay text in videos represent where, when and who of the happening events. A large number of techniques have been proposed by various researchers in the recent past to detect and localize the overlay text present in images and videos but there is still a lot to achieve in this field. Our aim is to build an efficient and accurate automatic overlay text detection and localization system.