Text Enhancement in Projected Imagery

There is great interest in improving the visual quality of projectedimagery. In particular, for image enhancement, we would assertthat text and non-text regions should be enhanced differently inseeking to maximize perceived quality, since the spatial and statis-tical characteristics of text and non-text images are quite distinct.In this paper, we present a text enhancement scheme based on anovel local dynamic range statistical thresholding. Given an inputimage, text-like regions are obtained on the basis of computing thelocal statistics of regions having a high dynamic range, allowing apixel-wise classification into text-like or background classes. Theactual enhancement is obtained via class-dependent Wiener filter-ing, with text-like regions sharpened more than the background.Experimental results on four challenging images show that the pro- posed scheme offers a better visual quality than projection with- out enhancement as well as a recent state-of-the-art enhancementmethod.