How to save a "duplicate" image

[wanderer]

Hi:
I have a burst of photos taken that I would like to keep.  One image in the set is flagged as a "duplicate" using the "visually identical, same file format" criteria.  They are not visually identical as the eyes and facial expression are slightly different.

How can I mark the "duplicate" as not a duplicate so that it isn't continually flagged when I add more photos to the database?

The photos are of a person, so understand they cannot be posted here.

[sybersitizen]

Maybe this: Open the built-in Duplicates category, select the files in question, right-click, and choose the Un-assign from current category function.

[Mario]

They are not visually identical as the eyes and facial expression are slightly different.

If you work with burst shots where each image in sequence only differs in a few pixels from the previous one, the "visually identical" detector will not work for you.

This detector does not compare image pixel-by-pixel, because this would take a very long time. Instead it looks at a very small rendition of the image (think: unsharp thumbnail) to determine the similarity of two images using math. And this method does not "see" that the eyes in the image are slightly different compared the other image.

How can I mark the "duplicate" as not a duplicate so that it isn't continually flagged when I add more photos to the database?

IMatch runs the duplicate check when new images come into the database. When it finds an image it considers a duplicate, it adds it to the Duplicates category and shows a notification. If does not show the notification of the image is already in the duplicates category.

But that happens only once, usually. Or do you still update the very similar images with an external application so IMatch rescans them and finds them again as dupes?

Anyway, if you work with burst shots often, just disable the automatic duplicate check (Edit menu > Preferences > Indexing) ("Don't check") or switch to the "binary identical" mode.

[wanderer]

Thank you @sybersitizen.
The "U" was the trick.  I looked for that command in the duplicate search/compare window but didn't think to back up to the category view. Still very much a noobie.   

Also, as I add folders to my test database, I seem to have re-indexed the folder by accident too. Good way to re-enforce the learning.

Mario:  Thank you.
The visual search capability is one of the tools that brought me to iMatch so want to keep it. 

Confusing to me, is that for my burst collections iMatch flagged only a single image as "duplicate" out of the group, and this was a non-sequential image.  The other intermediate images did not get flagged as duplicate(s) even though they looked visually similar. For most bursts the flagged image did look visually similar, but for this one the "flagged" image clearly had a different facial expression.  BTW, these are iPhone, so not a fast
"burst".

Is there a way to configure "burst" detection on indexing?  In these specific cases, a burst flag would seem more appropriate to me than a single "duplicate" flagged out of each burst group. While I wanted to keep the specific burst with the facial expression changes, most bursts needed more than one image deleted to clean up the collection.

Managing "file relations" is a bit above my skill level so far, so I am taking your advice by adding just a few folders into the database at a time as I work through the issues and my learning curve.

Thank you.