Caption Detection in News Based on Prewitt Operator
Shao Chen-xi · Electronic Technology · 2009
Caption in news video is one of the important sources for the realization of automatic retrieval,analysis and comprehension of videos since it implies lots of semantics inherently. On the basis of analyzing the characteristics of news captions,this paper proposes a news caption detection method based on Prewitt operator. The algorithm first makes gray-scale transformation of images and carries out edge detection based on Prewitt operator,finally,captions are extracted through regional exploration and captions merger. On test with different news video frames in experiment the accuracy rate 96.8% is obtained. The experimental results show that the approach proposed can well accomplish the task of detection news caption.