People of all generations are making more and more use of digital imaging systems in their daily lives. The image content rendered by these digital imaging systems largely differs in perceived quality depending on the system and its applications. To be able to optimize the experience of viewers of this content understanding and modeling perceived image quality is essential. Research on modeling image quality in a full-reference framework – where the original content can be used as a reference – is well established in literature. In many current applications, however, the perceived image quality needs to be modeled in a no-reference framework at real-time. As a consequence, the model needs to quantitatively predict perceived quality of a degraded image without being able to compare it to its original version, and has to achieve this with limited computational complexity in order to enable real-time application. Although human beings effortless judge image quality in a real-time no-reference framework, developing a model to simulate this perception is still an academic challenge partly due to our limited understanding of the human visual system. This thesis presents some achievements in designing no-reference objective quality metrics, which have the aim to automatically and quantitatively predict perceived image quality. Two different approaches are used. In one approach the perception of some specific image degradations is modeled. This approach is applied to the perception of blockiness and ringing, two degradations typically occurring as a consequence of signal compression. The resulting metrics are based on a two-steps framework: a first step, in which the artifacts are located and a second step, in which the local visibility of the artifact is estimated. Both components include aspects of human vision with which the reliability of the metrics in predicting perceived artifact annoyance is improved, while keeping the computational effort limited. In a second approach the overall perceived quality of images is predicted. An accurate and computationally efficient way to do so exists of combining a simplified feature extraction strategy – resulting in features based on aspects of the artifact specific metrics – with an adaptive neural network. After having trained the overall quality estimation system off-line, the metric can be very easily implemented in real-time devices. Whether the artifacts in an image attract the viewer’s attention also affect the viewer’s quality estimation. Hence, in a final study the improvement in quality prediction performance of various metrics by including visual attention is evaluated. In these metrics local quality information is weighted with the attention given locally by the averaged viewer. Results show that when using ground-truth attention obtained from eye-tracking recordings the degree to which the quality estimation is improved, depends on the type of metric and kind of image content.