Viewer context menu - add face detection

[ben]

FR
In the context menu of the viewer, please add the menu item "Detect faces and create face annotations"

Currently, if you go through images in the viewer and find an image without face annotations, you have to go back to the file window.


Thanks for considering this,
Ben

[Mario]

The way to do this: Click on the face icon in the annotation panel in the Viewer.
This is how this has worked since IMatch 5. Why do you go back into the File Window? See also the People Help topic which explains this and shows the icon to click.

[Jingo]

The way to do this: Click on the face icon in the annotation panel in the Viewer.
This is how this has worked since IMatch 5. Why do you go back into the File Window? See also the People Help topic which explains this and shows the icon to click.

I think Ben may be thinking that you have not run the AI Face Detect through the images to begin with and are just browsing images and find one with a face.  I wonder if we need an indication of some sort that detection has already been run on an image - something stored (and accessible) to indicate FD has already occurred and ask if we want to run it again on the image?  I know it doesn't add much to an image.. but if this would at least then tell the user that a manual face region will be needed since the AI couldn't locate one so running FD again won't be beneficial.

[Mario]

How is this in any way related to the original FR?
To many threads are currently hijacked or diluted because people seem to just post where they want...

[Jingo]

How is this in any way related to the original FR?
To many threads are currently hijacked or diluted because people seem to just post where they want...

I misunderstood your answer... you are saying you can ALREADY do this if you just click the face icon... I thought you were wondering WHY one would do this thus my discussion point.  My apologies....

[ben]

The way to do this: Click on the face icon in the annotation panel in the Viewer.
This is how this has worked since IMatch 5. Why do you go back into the File Window? See also the People Help topic which explains this and shows the icon to click.

I had the following scenario several times now.

• I first run face detection from the file window with normal mode (not optimized for small faces).
• Then when walking through all pictures in the viewer, there is an image with several faces. But only some of them have been detected by the face detection
• According to the help, by clicking the face annotation icon, face detection will be started in optimized mode, which would be good
• But, since the image already contains face annotations, instead of running the face detection only a default face rectangle is shown

Thus, i can either create face annotations manually for the rest of the faces, or i need to go back to the file window to run the face detection (optimized for small faces) again.
Or, am i still missing something?


So, maybe we can change the initial FR to:
In the viewer, allow to (re-) run face detection manually and by user selected mode.

Ben

[Mario]

There is 'additive' face detection. This would just duplicate already detected faces.
Or I would have to add overlapping detection and weed out newly detected faces. But this will cause problems in case faces are close together or face annotations overlap.

You can run face detection in the Viewer again by deleting existing face annotations first.
Or just click the button to add a face rectangle and position it manually. This works 100% correct all the time and is also quick.