As I mentioned in my previous post, I’m dedicating some time to dive in to Lightroom’s facial recognition feature. Initially, I wasn’t having a lot of luck actually using the feature itself, so I branched out a bit to do some tangential exploration.
This week, I wanted to pass along the note that Lightroom stores the facial recognition regions and names in the XMP sidecar files. If you are automatically generating XMP files, after you add the facial region and the name inside of Lightroom, the XMP file will reflect the area and the name in the Regions section. If you are not automatically generating them, you can use ⌘S to save the facial recognition metadata to the sidecar file for the current image.
If you add a region without a name or if you remove the name, Lightroom will still store the region in the XMP file, even if there is no associated name. This is helpful since automatically finding the face regions in your catalog can be a resources intensive process to go through more than once.
Since I’m still waiting for the preview creation and facial scan on my primary catalog and have tons of free time, I also went through the exercise of creating a new catalog and importing images with the XMP sidecar files to verify that the region came back intact, which they did.
I still rely on XMP files for some of my archives, so if I ever do get the facial recognition feature working and decide to include it in my workflow, I’m glad that all the work that I would hypothetically to do tag all the faces in my catalog will be maintained when my I push my images out to the archive pasture.