ALE
Image Processing Software Deblurring, Anti-aliasing, and Superresolution. Local Operation localhost 5393119533 |
[ Up | Monte Carlo ]
Error functions are used to determine whether a transformation is "good" or not, and in particular, whether one transformation is better or worse than another. In particular, a smaller error indicates a better transformation.
ALE defines error functions for each pixel of the accumulated image, and also a comprehensive error function that summarizes the error over all pixels.
Since differences in exposure may alter the absolute values of pixels even in images that are perfectly aligned, ALE usually normalizes the images so that overall image intensity matches. This normalization step does not affect the pixel values contributing to the merged image, but it does affect the values with which the per-pixel error function is calculated. (The normalization step is skipped when --extend or --follow is specified.)
The per-pixel error function for pixel (i, j) in the accumulated image A, frame B, and transformation T, is:
p(i, j, A, B, T) = Abs( A(i, j) - B(T_inverse(i, j)) )metric_exponent
where Abs() is the absolute value function and B(T_inverse(i, j)) is the overlapping value described in the merging section. If not specified by the --metric option, metric_exponent defaults to 2.0 for ALE versions 0.1.0 and later, or 1.0 for version 0.0.0.
In addition to the per-pixel error function, a per-pixel maximum error estimator is calculated, as follows:
p_max(i, j, A, B, T) = Max( A(i, j), B(T_inverse(i, j)) )metric_exponent
If we consider Sum[] to provide the sum over whatever subset of pixels we are sampling (including possibly the set of all pixels), then the comprehensive error function, for accumulated image A, frame B, and transformation T, is:
E(A, B, T) = (Sum [ p(i, j, A, B, T) ] / Sum [ p_max(i, j, A, B, T) ])(1/metric_exponent)
Verbatim copying and distribution of this entire article is permitted in any medium, provided this notice is preserved.