An important aspect of image and print quality is the existence of artifacts, such as compression or print artifacts. A general perceptual masking model, that describes the perceptual severity of artifacts on general background, could have been used to extract specific artifact detectors. However, currently general models are not mature enough to provide print artifact detectors for commercial print quality control application. Consequently we propose to employ machine learning techniques to learn a specific model for each print artifact based on a relevant set of features. We used the approach to develop two print artifact detectors. While the proposed approach was developed for print quality purpose, the method is general and can be used for learning automatic evaluators for image defects and quality degradation as well.
Hila Nachlieli, Hadas Kogan, Morad Awad, Doron Shaked, Smadar Shiffman, "Learning Print Artifact Detectors" in Proc. IS&T CGIV 2012 6th European Conf. on Colour in Graphics, Imaging, and Vision, 2012, pp 81 - 85, https://doi.org/10.2352/CGIV.2012.6.1.art00015