Artifact reduction in multiscale contrast enhancement for digital radiography

Nonlinear multiscale detail and contrast manipulation of digital radiographic images produces artefacts that reduce the fidelity of the image and have a distracting effect on the observer. These artefacts are most prominent on the borders of highly exposed parts of the image, like the skin line region. In this paper we present a novel method for multiscale image detail, local and global contrast enhancement that adapts to local exposure level and thus reduces the artefacts´ saliency. Proposed algorithm was incorporated into the digital radiography framework and compared to a reference one without adaptation to intensity. Evaluation on 60 images of different anatomies from different patients showed that the proposed algorithm performs better or equal to the reference one in 93% of the cases without manual adjustment of the parameters.