In this paper we propose an FR attention driven foveated video quality (AFViQ) metric. A perceptual based foveal imaging model consisting of an improved video attention model and an advanced fixation prediction approach by simulating the essential functionality of eye movement; ii) an accurate video quality metric by integrating the foveal imaging model into wavelet-based distortion visibility for quality assessment. Extensive experiments conducted on publicly available eye-tracking and video quality assessment databases have demonstrated promising performance of the proposed video attention model, fixation prediction approach, and the perceptual quality metric.
You are here: Home / ieee projects 2014 / Foveated Video Quality Metric based on Wavelet Distortion Visibility