Title :
A High-Performance Videotext Detection Algorithm
Author :
Hsia, Shih-Chang ; Ho, Cheng-Nan
Author_Institution :
Dept. of Electron. Eng., Nat. Yunlin Univ. of Sci. & Technol., Yunlin, Taiwan
Abstract :
In this paper, we propose adaptive texts detection and localization algorithms based on an edge extraction approach. The main techniques include edge filtering, adaptive threshold, the morphological operator and pixel correlation. With block-based processing, the rough texts can be detected by edge strength. The morphological operation is employed to refine the text region and to remove discrete non-text marks as well. The final text region can be accurately localized with a rectangle by horizontal differential approach. Simulations show that the proposed method can attain higher accuracy and lower false detection compared with the recent top paper.
Keywords :
edge detection; image segmentation; text analysis; viewdata; adaptive texts detection algorithms; adaptive texts localization algorithms; adaptive threshold; block-based processing; discrete nontext marks; edge extraction approach; edge filtering; edge strength; false detection; high-performance videotext detection algorithm; horizontal differential approach; morphological operator; pixel correlation; text region; Algorithm design and analysis; Detection algorithms; Image edge detection; Information filters; Signal processing algorithms; Streaming media; Rober filter; morphological operation; text detection; videotext;
Conference_Titel :
Intelligent Information Hiding and Multimedia Signal Processing (IIH-MSP), 2012 Eighth International Conference on
Conference_Location :
Piraeus
Print_ISBN :
978-1-4673-1741-2
DOI :
10.1109/IIH-MSP.2012.65