A novel processing method of the subjective quality evaluation results of stereo video

摘要

Subjective quality assessment of stereo video has been the standard that can judge and amend whether every link in the field of stereo imaging technology is reasonable or not because of its true and reliable results. However, many unstable factors in subjective experiments will bring abnormal values to experiment results. Therefore, to get more reliable experimental results, a reasonable data-processing method is designed in this paper to analyze the final results. The method includes three aspects: elimination of abnormal cases, remove of abnormal values and mean value treatment. Firstly, normality of the data sample is analyzed according to probability statistics. By statistical measurement parameters, the author calculates the corresponding measurement parameter and eliminates the abnormal cases. Then, abnormal values are removed by improved t test method. Finally, mean value treatment is carried out. The experimental results show that this method is obviously better than the traditional Grubbs test and it is extremely applicable.