Thanks for taking time to ponder this question. First a short introduction of the scenario. Recently retired, I am now embarking on capturing my family history. This will span from approximtely 1930 to the present and involve about 3-400 different people. It involves approximately 50K photos.
It took me about a week for the program to complete the facial recognition without crashing. I intentionally stopped the program and optimized the catalog a couple times per day and the crashes seemed to stop. Prior to this I would get "out of memory" errors and lose my work.
As such, the facial recognition has been a key tool to help me create EVENTS, which is the primary way in which I want our family narrative created. Just having a list of every time that person appears in a photo seems useless to me. But being able to quickly identify "who was where when and doing what with whom" is important. Sorry that was an akward way to say it! For example, "Grandpa's trip across America" in 1930 would be an Event I wish to preserve and the facial recognition goes a long way in helping me put the character names together.
NOW THE QUESTIONS
1. The catalog is now large, has crashed a few times during facial recognition and has been difficult to restore properly. In order to more safely save the work, should I split it into pieces (probably by decades or years) ?
* Technical Data: Late 2013 iMac 21.5" w 8 gb ram and 250 SSD HD. The catalog is stored on an external 1TB SSD. The backups are on an external 5400 rpm HD. A full backup takes quite a while to complete. Optimization takes only 15 minutes or less.
2. Can the facial recognition database be used universally over every catalog? In other words, is there any way, if I split the catalog into decades for example, that I will not have to begin again with facial recognition?
Thank you all so very much. I am new to this, I'm an old school film photographer totally enamored and intimidated by the power of this awesome technology!