A modified unsharp-masking technique for image contrast enhancement

Unsharp masking is a popular and simple technique for contrast enhancement and sharpening in digital images. The basic idea in this technique is to emphasize edges and discontinuities in the image by adding the edge information back to the original image. In order to achieve higher contrast level in the processed image, the edge information can be scaled prior addition to the original image. However, there are no guidelines on how to specify the scaling values. In this paper, we propose a novel technique to scale the edge information automatically. The underlying principle in the proposed technique is to apply the well-known histogram equalization technique on the edge histogram of the image. Equalizing the edge histogram is performed on selected regions of the edge histogram in order to reduce noise amplification in smooth regions and edge ringing artifacts. The proposed technique has been tested on a large set of images qualitatively and quantitatively and is proven to produce satisfactory results.