This chapter provides a comprehensive overview of quality assessment and perception models for images and videos, focusing on key concepts such as subjective and objective quality assessment frameworks, Just-Noticeable-Distortion (JND), Salient Object Detection (SOD). Subjective and objective quality assessment methods are examined, with subjective approaches like Mean Opinion Score (MOS) serving as benchmarks, while objective methods—Full-Reference (FR), Reduced-Reference (RR), and No-Reference (NR)—leverage metrics such as PSNR and SSIM, alongside deep learning models for enhanced performance. JND is introduced as a critical perceptual threshold in the human visual system (HVS), distinguishing perceptible distortions from imperceptible ones to optimize quality assessment models. The discussion on SOD covers both image and video domains. This chapter also reviews JND datasets and models in pixel and frequency domains, emphasizing their applications in image/video coding and quality assessment. By integrating perceptual models with advanced computational techniques, this chapter underscores the importance of aligning quality assessment with human visual perception for multimedia applications.