Photo.app helped me curate and catalog GBs of photos from twenty+ years of archives. It sped up the family project I was working on, raised my expectations for how Enterprise photo asset management can work, and compelled me to reconsider ways to take advantage of organizing the rest of my work and personal archives.
During import, Photo.app automated tons of organization by geo- and date-tagging items. Facial recognition was decent during initial post-processing. With training, it got very good and helped me build & populate People collections for current and future projects. With each new batch of photos I threw at it, People collections grew, and the app would present additional photos to review for inclusion in individual People collections.
Good automated guesses + smart confirmation request + ongoing opportunities to manually update collection inclusions and exclusions is a strong combination for a consumer app.
The app asks for help and confirmation appropriately. It wondered if my brother, Ed, was also Ed in photos from when he was 7, 12, and 19. Manually confirming a few photos for a collection adds those items explicitly and helps the app find more from the rest of the library. Here I confirmed 5 photos of myself and Photo.app added those and 59 more — impressively seeing through multiple angles, ages, and looks.
Even when it’s wrong, it’s often close. Like asking if my nieces and nephews are their mother or father. When it’s really wrong, the suggestions are usually reasonable and sometimes funny.
Update: Successes with Photo.app during this project compelled me to plot out getting archival documents, video, images, audio, and code catalogued, discoverable, and accessible for other personal and professional projects.