The volume of UGC and social-media video is enormous, however, the number of high-quality or broadcast quality material is only a small portion of the original set. In COGNITUS, algorithms that can measure the quality of media in terms of its visual aesthetics and attention are developed. These metrics are closer to the human perception of media than Mean Square Error (MSE) or Peak Signal-to-Noise Ratio (PSNR) metrics. Thus, aesthetic and attention metrics become fundamental tools to the professionals so as to first filter content that has little interest in terms of perceived quality. These QoE metrics include attention (focus, faces, faces area, smiles, saliency, focus difference, motion magnitude) and aesthetics (colour moments, colour ratio, luminance, luminance standard deviation, edges strength, edges orientation, colour diversity, rule of thirds, foreground area, entropy, shakiness, shadow area, camera movement).

COGNITUS provides the tools that automatically inspect the visual quality of such videos and selects which low quality videos can be removed. Such tools become a critical resource for many media professionals who are guided by QoE.