Curating 20 years worth of photos

A few years ago, Whatsapp didn't let you choose which photos to download, it download them all. As a result, most people's photo library are filled with a massive amount of horrible memes sent in groups which you were to socially aware to leave.

I checked, and I had more than 4000 photos in this category.

As part of my technological declutter, I wanted to clean up my photo library. It was not the largest, but not the smallest, sitting at around ~40 GB, or 35000 photos.

Most of the photos were not mine, I had copied family photos to my own library thinking that I wanted to have full archival of things I might want to see at some point in the future.

Since then, my approach to photo management has changed. I don't want to have more than 3 or 4 photos per single day. I don't want photos of places I haven't been personally if there are no people in them. I don't want memes. I don't want screenshots. I don't want 10 copies of the same photo from slightly different angles. I don't know who invented bursts, but you are on my list.

Last weekend I took it upon my self to clean up and I have to admit that it took a lot more effort than what I care to admit. It's done: now my library consists of measly ~5000 photos.

The approach I finally took is as follows:

• Generate a thumbnail of all photos and vibe code a dirty webpage to select photos to keep and delete. I want the tools to make most of the decision by me, but I want final say.
• Czkawka for deleting duplicate photos. Quite useful, but had to play with thresholds.
• Wrote a script to detect text, 95% photos with text were things that I wanted to delete (mostly memes).
• Wrote a script to detect faces. Most photos without faces I wanted to delete but this had a lot more false-positives. Useful, but not as reliable as the other two.

Beyond the above, I manually reviewed all the remaining and deleted a bit more. Quite happy with the results. Happy deleting!