ENHANCEMENT OF SCREEN CONTENT IMAGES BY JUST NOTICEABLE DIFFERENCE MODEL

Abstract

Now we can observe the rapid development of cloud computing services are influenced by the transmission of screen content images (SCI). While transmitting the images they are going to be compressed to decrease the computational time. The visual quality of screen content images will be reduced at this time. To overcome this we introduce a new just noticeable difference (JND) model for the enhancement of screen content images before transmitting through remote systems. JND model adopts a local parametric edge model to generate edge profile, which is an enhanced screen content image. This parametric edge model predicts the luminance, contrast and sensitivity of an edge profile by making the use of luminance adaptation, contrast masking and structural sensitivity algorithms. By using these parameters proposed model evaluate the visibility threshold. This just noticeable difference model estimates on and near edge pixel transitions. If we see the visual quality of screen content images exhibited by the parametric edge model outperforms the existing JND models.