Visual Attention Complexity Of Scene

With the development of information technology, the semantics-based video qualitymetrics considering the content of scene were proposed, since the pixel-based video qualitymetrics could not meet the requirement. The visual attention complexity of a scene is regardedas a feature describing many information of the content in a video sequence, whichmeanwhile plays an important role in various applications to estimate the perceptual quality,information retrieval and so on.In the beginning, this thesis presents the overview of research in the field of visualattention complexity, including the image and video visual attention complexity and proposedperceptual oriented video quality assessment methods.In the following part, the hypothesis about visual attention complexity is proposed. Thecomplex video sequence in the view of visual attention should contain a large quantity of“informative” objects on the scene. So, the visual attention complexity(VAC) indicatorextraction from a video sequence is conducted by information theory on saliency mapgenerated from computational visual attention model. The VAC indicator’s performance isanalyzed with the ground truth from an eyetrack database of IRCCyN/IVC.The proposed VAC indicator is applied in video quality estimation method which iswidely used in video transmission or compression system. In addition to VAC indicator,spatial and temporal information from original video sequence and Bitrate or PSNR fromcompression video compose the set of elements for the quality estimation. The objectivevideo quality estimation model is based on a machine learning algorithm and tested on aH.264compressed video database of IRCCyN/IVC. All proposed models are verifed byanother H.264compressed video database of IRCCyN/IVC. The quality scores predicted byproposed models have a high correlation coefficient to the subjective quality scores.