This is an old revision of the document!
Compare
Usage
Compares images with a given metric. If the number of images in the two inputs match. a 1-to-1 comparison will be performed. If the numbers dont match every images in input 1 will be compared to every image in input 2. Thereby the number of images in input 1 needs to be bigger number of images in input 2.
Parameters
Different modes for comparing images can be chosen.
Modes | Description |
---|---|
CCC | Cross Correlation Coefficient is used as metric. normalize: enables normalization of the images before calculation. This should be done if the images are not normalized up to here! radius: radius of a centered circular mask which is used before comparing the images. If the value is 0 no mask is used. |
covariance | Correlation Variance Ratio of the images |
Fourier ring correlation | Fourier Ring Correlation of the images is used as comparison metric |
normalized mean square error | Normalized mean square error is used as comparison metric |
sum of absolute differences | Sum of Absolute Differences is used as comparison metric |
Inputs
input1: stack of images or single image (stack needs to contain equal or more images than input2)
input2: stack of images or single image (stack needs to contain equal or less images than input1)
Output
output: unchanged image stack from image1 containing a value for the metric (in case of unequal image numbers the best value) and the image number of the image from input2 giving the highest metric value
Written Header Values:
- ccc-/nmse-/phash-/sad-/frc-/cov-/mui-/pHashPeakHeight: Comparison value of the two images from the inputs on that position with chosen metric
- ccc-/nmse-/phash-/sad-/frc-/cov-/mui-/pHashPeakReferenceIndex: ImageID of the image from input2 that gives the highest metric value (only written if number of images in input1 is bigger than number of images in input2)