Similar to my Google Photos experiment, I set out to organize the same set of photos both in Amazon Photos and Memorable.
I uploaded 48,993 photos to Amazon and Memorable through their respective desktop apps.
After uploading everything, I took note of Amazon’s default numbers before merge.
I then scrolled through the entire people list, looking to combine any of the top four people.
Once merged, Amazon Photos’ numbers were higher than Google Photos for each main person.
Memorable uses Amazon’s Rekognition service, the same technology used by Amazon Photos to automatically organize pictures by person.
I figured out a way to overcome Amazon’s limits and produce better results by adding a user experience (UX) layer on top of the data from their service.
By adding a similarity threshold slider, Memorable can find more photos of the same person.
Memorable improves upon Google’s and Amazon’s merge process by eliminating the need for a name, showing multiple possible matches at the same time, and finding smaller groups of the same person.