Face
Detection

This was a one-night build.

We have a bunch of old family photos, but a lot of them are just landscapes—trees, lakes, streets. Not super exciting. What we actually care about are the faces—people doing things, interacting, being themselves.

I used Python with OpenCV and their built-in CascadeClassifiers. The starting point was this tutorial.

My steps:

• Detect faces
  First pass was just running detection, but results were noisy. I improved accuracy by running multiple face detection models and checking if at least half agreed—if so, we mark it as a “face photo.”
• Walk through folders and sort
  The script walks through a directory tree, processes each image, and moves anything with a face into a separate folder.
• Package into an executable
  Final step was bundling it up. I had to tweak the config to include all the CascadeClassifier files, but nothing too complex.

That’s the app. Just a quick utility to surface the images we actually care about.