In the past decades, researchers in video coding and image processing pursue for high fidelity of image/video communication through limited transmission bandwidth. The system performance is often measured with signal to noise ratio (SNR), that is, perfectly image/video reconstruction is the goal of these systems. However, the final receivers of image systems are actually human vision systems. System design with SNR as the target is often overkilled and sometime ill-posed. In recently years, a raising field is perceptual image/video processing, where human perception models are taken into consideration to further improve the system performance. In this talk, we will provide examples in video coding and image processing for display devices: with perception model, more than 30% of the bitrate can be saved while the perception quality is well kept; with more understandings in human perception, we can further reduce the cost and power consumption, and increase the quality of display devices.