Once DigiKam completes face detection / identification on group pictures or public events for instance, we may end up with dozens even hundreads of results. Manually clicking on each and every Ignore or Remove button (when only one person is significant) is extremely cumbersome and this is made worst by the fact the program need to write (and read it again since there seem to be a refresh of the zones in the picture?) to the file for each button click.

As mentioned in the directives, I don't want to hint too much at any good way of presenting this in the UI but I think it may help to understand my idea. Globally, what I would imagine is to have a panel which presents a list of the faces identified in the picture, similar to what we see in the left people pane. They could be grouped by Confirmed and Unconfirmed and allow to select multiple (ctrl, shift, checkbox click) and then, click on either Remove or Ignore.

While we're at it, this could also be an opportunity for bulk assigning by allowing to match the faces directly in that UI and apply every one at once by clicking on a Confirm button so that the file is written only once.