2. Example: Thresholded division of elements

Division of two elements is commonly used to normalise one element to another. Without thresholding of low background values this can produce a rather unattractive image. See also Calculator.

calc_img_bad = calc_img_z66 / calc_img_p31

In this example we use the calculator tool to simultaneously threshold and divide two elements.

calc_img_div = calc_img_z66 / calc_img_p31

  1. Determine the background value of the divisor image.

    In this example a value of 100 sufficed.

  2. Using the calculator tool perform the operation.

    Enter if P31 > 100 then Zn66 / P31 else 0 into the Formula box.

    The first part of the if/then/else masks the data so only values above the threshold are operated on. The second part performs the division while the last part sets unmasked data to 0.

  3. Enter the new name, and click Apply.

    Choose a unique name to prevent overwriting data.

3. Example: Image Segmentation

Selection of a region of interest of structural feature can be performed using segmentation. In this example we will use multi-level kmeans to select certain features from Fe56 of a tissue-micro-array core.

calc_img_k_og

  1. Create the segmented image mask.

    Enter segment(Fe56, kmeans(Fe56, 3)) into the Formula box. Here we are using 3 level kmeans clustering.

calc_img_k_mask

  1. Enter masked into the Name box, and click Apply.

    You can choose any name will work.

  2. Select data from the original.

    Enter mask(Fe56, masked == 3) into the Formula box. The selected data is limited to regions where k==3.

calc_img_k_select

  1. Enter the new name, and click Apply.

    Choose a unique name to prevent overwriting data.