Organization and retrieval of continuous media
MULTIMEDIA '00 Proceedings of the 2000 ACM workshops on Multimedia
Automatic News Video Caption Extraction and Recognition
IDEAL '00 Proceedings of the Second International Conference on Intelligent Data Engineering and Automated Learning, Data Mining, Financial Engineering, and Intelligent Agents
Hi-index | 0.00 |
In accumulating and retrieving contents of video data, it is necessary to extract the telop (video caption) and flip characters efficiently and accurately, because they summarize the video content concisely. The purpose of this study is to automatically extract telop and flip characters. It starts from the extraction of stable frame sections including the telop and flip characters. In this extraction, we propose the detection of the telop disappearing frames in addition to the telop appearing frames. Next process is the character region extraction and here we propose application of local line density to discriminate telop characters from other elements such as lines and symbols. Final process is the telop character extraction and here we show the effectiveness of the floating adaptive three level thresholding (FATLT) which thresholds the image intensity into three levels at first and finally binarizes the character regions precisely even in low contrast, taking topological relation between characters and their background into consideration.