Implementation of Text Extraction From Video Using Morphology and DWT

Abstract

Video has one of the most popular media for entertainment, study and types delivered through the internet, wireless network, broadcast which deals with the video content analysis and retrieval. The content based retrieval of image and video databases is an important application due to rapid proliferation of digital video data on the Internet and corporate intranets. Text which is extracted from video either embedded or superimposed within video frames is very useful for describing the contents of the frames, it enables both keyword and free-text based search from internet that find out the any contained display in the video. The algorithm performance on the basis of text localization and false positive rate has improved for different types of video. The overall accuracy of this methodology is high than that of any other methods. The advantage of this algorithm is that it minimise processing time.