Intel IPP functions described in this section compute the universal image quality index [Wang02] that may be used as image and video quality distortion measure. It is mathematically defined by modeling the image distortion relative to the reference image as a combination of three factors: loss of correlation, luminance distortion, and contrast distortion.
If two images
gare considered as a matrices with
Nrows containing pixel values
i,j], respectively (0 ≥
M, 0 ≥
N), the universal image quality index
Qmay be calculated as a product of three components:
The first component is the correlation coefficient, which measures the degree of linear correlation between images
g. It varies in the range [-1, 1]. The best value 1 is obtained when
gare linearly related, which means that
+bfor all possible values of
j. The second component, with a value range of [0, 1], measures how close the mean luminance is between images. Since
can be considered as estimates of the contrast of
g, the third component measures how similar the contrasts of the images are. The value range for this component is also [0, 1].
The range of values for the index
Qis [-1, 1]. The best value 1 is achieved if and only if the images are identical.