| 摘要 |
A system and method for detecting text edges in digital document images utilizes filters (202, 218, 222, 226) that process the luminance and/or chrominance values within a selected region (110) of the image (108) to determine whether a given pixel (106) of the selected image region is a text edge. The filters are configured to determine the gradient (202), the curvature (218), the maximum luminance value (222), and/or the maximum chrominance value (226) for the luminance values within the selected region of the document image. Two or more of these indicators are used to detect text edges in the document image. <IMAGE> |
