Date of Publication :7th May 2016

Abstract: Subtitle extraction from video sequences finds numerous useful applications in video classification. The complex background, illumination and font size of text present in the video makes subtitles extraction extremely challenging. Text in the video are classified as scene text and graphics text. In this paper, character segmentation of graphics text is implemented using morphological operations. Compared to word segmentation in video subtitles, single character segmentation can achieve higher precision because of the simple background. In this paper, the English text present in subtitle of the video is considered for processing. All the character segmentation steps are implemented in Microsoft Visual Studio 2010.

Reference :