Text extraction algorithm under background image using wavelet transforms

With the growing number of digital multimedia libraries, the need to efficiently index multimedia information is increasing, detecting and extracting the text information from images plays an important part in images indexing based on content. In the paper, a new text extraction algorithm under background image based on two-dimensional wavelet transforms is proposed. For the algorithm, firstly the image is transformed into the wavelet domain and then a sliding window is set to scan high frequency sub-bands, through computing the wavelet texture features of the image in the sliding window, k-means clustering algorithm is used to classify the image into text area, simple background area and complex background area. Finally mathematical morphology operations are applied on the text area to locate the text positions exactly. The experimental result shows that the algorithm can extract text information with different languages, fonts, sizes and ways of arrangement from the background image exactly.