For the annotation of images with different image sizes, annotating hyperstack is not straightforward.
One could give the option to annotate one image at a time by annotating from a directory on disk.
Ask for "stack" or "from disk" in a dedicated option window before generating the GUI ?
Rather for the article but currently all post-processing is done in KNIME.
One example in ImageJ/Fiji would be appreciated, try the sunburst plot in ImageJ/Fiji with JFreeChart ?
Suggested by reviewers (see article page on F1000)
With multiple classes, hard to figure out which shortcut for which category
Either use a pop up balloon to give a hint about the shortcut or directly put it in the button label
Continuing an annotation from an existing table could be made easier by giving the possibility to generate the category choices from an existing table.
Give this option in the pre-annotation window ?
Button "Load from file" or "...from current table"
Currently annotating the same image or ROI will result in duplicate entries in the table.
Give the option to replace existing entries.
This means checking some column values to see if they contain the value which is about to be appended.
The branch overwrite-bruteforce-search adds a checkbox "Overwriting existing" and a function to find existing rows provided a folder, image and optional slice and roi identifier.
Problem: when measurement is selected the analyzer automatically adds to the result table
From there possible solutions:
make the analyzer write to another table with a unique ro, read this row into a dictionary and close the table
check if Measure: read last row, delete last row, replace existing row
modify the Analyzer class so that it returns the measurement into a variable (but then not available for current version of ImageJ/Fiji)