Research Of Video Quality Objective Assessment Algorithms

With the rapid development of networked video applications,such as video on demand,digital television,video conference,Internet streaming video and video surveillance,digital video is getting into our daily lives.From the front video capture device to the end user,in order to ensure the quality of video service and experience of user,we need to assess the quality of the video image from the video processing and transform and other terms.Since human eye tends to be the ultimate recipient of the video images,an ordinary people with some life experience can accurately evaluate the clarity of the video image,video coherence,and vividness of color and image saturation.However,for the huge amount of video data,subjective video quality assessment has been powerless.Objective video quality assessment methods which utilize mathematical models and computer programs to automate video quality evaluation have become a hot topic in the field of video quality assessment.Objective video quality assessment algorithms according to the dependence on raw video can be divided into full-reference(FR),reduced-reference(RR)and no-reference(NR)video quality assessment algorithm.This paper focuses on the full reference and none-reference video quality assessment algorithm,and the main work is divided into three parts: 1.According to the study on the video spatial and temporal characteristics of visual perception,an improved full reference video quality assessment algorithm-video quality assessment algorithm based on similarity gradient of spatial and temporal domain is proposed in this paper,which has better performance and lower evaluation algorithm complexity,compared with the current international full-reference video quality assessment algorithms.2.Several state-of-the-art universal no-reference IQA algorithms are reviewed in this paper.Firstly,the features extraction methods and quality assessment principle of each algorithm are introduced respectively.Then these algorithms are simulated on the LIVE image quality database.Lastly,the performances of the algorithms are compared quantitatively and qualitatively,and the advantages and disadvantages of these algorithms are analyzed.3.The common spatial and temporal distortions in video surveillance system are researched,and then No-reference quality assessment algorithms are applied to video surveillance system.At last the objective quality results are compared with the subjective quality results.