Intensity-based gain adaptive unsharp masking for image contrast enhancement

With the increasing application of image processing techniques in various areas, contrast enhancement still remains as a core pre-processing element that critically affects the performance of subsequent procedures. Among the available enhancement methods, unsharp masking is one of the efficient approaches that could produce satisfactory results on a wide range of digital images. However, properly setting the unsharp masking algorithmic parameters is a challenging task. In this work, a gain adaptive scheme is proposed where enhancements are focused on the mid-range intensity regions. This approach effectively makes use of a Gaussian kernel to adjust the enhancement gain and reduces the probability of over-range artefact occurrences. Experiments are conducted on a collection of public available digital colour images and in-house captured real world images. Satisfactory test results have illustrated the performance of the proposed method.