project
Photo organizer with face recognition
Python tool that detects faces and clusters event photos by person, replacing manual sorting.
A command-line tool that takes a folder of event photos and automatically groups images by detected person.
The problem was straightforward: after an event with hundreds of photos, manually sorting which images each guest appeared in took hours. The tool handles that in minutes.
how it works
Each photo is processed by OpenCV to detect faces. The face_recognition library (built on dlib) extracts facial embeddings and compares them against a set of reference photos. Images are then organized into folders by person.
what I learned
This was my first project with computer vision. I learned about the precision trade-offs of face_recognition under different lighting conditions and angles, and how the quality of reference photos directly affects results.