Abstract
Caption graphically superimposed in news video frames can provide important indexing information. The automatic extraction and recognition of news captions can be of great help in querying topics of interest in a digital news video library. To develop such a system for Chinese news video, we present algorithms for detection, extraction, binarization and recognition of Chinese video captions. Experimental results show that our caption processing scheme is e ective and robust and signi cantly improves video caption OCR results.
Cite
CITATION STYLE
Gao, X., & Tang, X. (2000). Automatic news video caption extraction and recognition. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1983, pp. 425–430). Springer Verlag. https://doi.org/10.1007/3-540-44491-2_62
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
