Study On Image Quality Assessment For Inverse Halftoning

The traditional method for image quality assessment aims at the image of compressing and video coding. But there are many differences between inverse halftoning and compressing. So the traditional method can not be applied into inverse halftoning image directly.On the basic of the advantage and the disadvantage of the traditional method, this paper analyzed the artificial texture and the noise which it is sensitive by human visual system. In the high-frequency of the inverse halftoning image, the indicators such as Gradient, Entropy and Inverse Deficit Moment has been created by the Gray Level Co-occurrence Matrix in order to reflect the edge area's quality. In the low-frequency, the Angle Second Moment has been extracted; it can reflect the smooth area's quality. The experimental results have shown that the algorithm is effective and adapts to inverse halftoning image.According to the color space model, the color inverse halftoning image's noise and artificial texture be kept into the brightness channel which in the HSI color space by linear transformation. The indicators of the artificial texture and the noise for color inverse halftoning image have been proposed in the brightness channel.