Text extraction in video image based on Wavelet Modulus Maximum

Text extracting in video image is a valuable task, because video text includes much useful information. In this paper, a new algorithm is proposed for text extracting. Use Wavelet Modulus Maximum to detect artificial text area in video image. Transform the image with wavelet and acquired the modulus maximum points. These points constitute the feature image. In the feature image extracts the feature to original and statistic feature to division the text area and the background. The experiment expresses that the result of the algorithm is better than other algorithms that are referred in the paper.