Research on DVD subtitles extraction and text automatic recognition technology
Liming Liu · Journal of Tianjin University of Technology and Education · 2009
To archive and manage DVD contents,it is necessary to process subtitles text.In order to get plain text data from subtitles,the characters in subtitle pictures have to be extracted and recognized.Based on researching and analyzing of DVD file organization mechanism and data storage structure of subtitle streams,the interfaces of accessing subtitles,private stream 1 and VOB files are explained.The method of separating and extracting English subtitles is presented in this paper.Five steps are used to separate DVD subtitle stream,extract subtitle pictures,split characters,and train samples recognize text.An implementation of this method can get English subtitle text automatically after simple training.