QUANTITATIVE COMPARISON OF IMAGES AND TRANSFORMS

Many different measures are currently used for quantifying the differe nce between pairs of images. Those in common use are surveyed, their i nterrelationships are made clear and their dependence on transform ban d-width pointed out. Adaptations to image transforms, with complex dat a, are given. For comparing simulated with experimental images, the ch i(2) difference is recommended as allowing a simple prediction of the likelihood of obtaining at random a difference as great as that actual ly obtained. In particular, the chi(2) difference in Fourier space is shown to be the only interpretable difference measure available in cas es when the image noise is not normally distributed or when - as is no rmally the case - the noise in neighbouring image pixels is not indepe ndent.