PyFaces
AI-powered smart assistant to classify faces in massive photo batches.
The problem
The immense time lost in visual sorting. In a graduation with 30 students and 700 photos, the photographer had to review the same gallery 30 times. This consumed entire days of exhausting post-production work.
What changed
I developed a Python engine based on facial recognition, packaged in Docker for frictionless installation. Given 3 reference photos, the algorithm scans the entire batch, compares the numerical proximity of the faces, and extracts all matches to a new folder.
The result
A monumental reduction in delivery times. Processing and identifying people in a batch of 700 images went from taking 3 to 4 days, to being completed autonomously in just 2 hours.
Background
The project was born from the real need of a former teacher and event photographer. After covering graduations or weddings, he faced the tedious manual task of separating and grouping photos by person to deliver them to his clients.
Technical details
Technologies
My role
Solo Software Developer
Does your company have a similar problem?
Tell me on WhatsApp and I'll tell you how we'd solve it.