Perceived visual quality metric based on error spread and contrast

This paper presents a new objective visual quality metric based on the measurement of error spread and isotropic local contrast of the image. It is shown that error spread is a reasonable measure for perceived visual quality while the image contrast is a content indicator for different images. The performance of the proposed method is shown to be superior compared to the 9 proponents in VQEG Phase I test dataset and another recently reported method in terms of prediction accuracy, monotonicity and consistency. In addition, the new method is very simple and does not need to tune any parameters.