Predict the MOS and the PSNR by the neural network

In order to change the judgment of the human eye by an objective measure, we present in this paper the new method to measure the quality of the video. This latter predicts the mean opinion score (MOS) and the peak signal to noise ratio (PSNR) by providing height parameters extracted from original and coded videos. The eight parameters that are used are: the average of DFT differences, the standard deviation of DFT differences, the average of DCT differences, the standard deviation of DCT differences, the variance of energy of color, the luminance Y, the chrominance U and the chrominance V. The results we obtained for the correlation show a percentage of 99.58% on training sets and 96.4% on the testing sets. These results compare very favorably with the results obtained with other methods, see examples used in [1]-[12].