Match the face to the frame.
The Face Shape Analyzer classifies a shopper's face shape from their webcam and surfaces frame recommendations from the brand's catalog matched to it — a quiz-style flow that ends in better fits.
From landmarks to a shortlist.
Shoppers facing a wall of 400 frames bounce. The Face Shape Analyzer turns a vast catalog into a short, personal shortlist — anchored to the geometry of the shopper's actual face.
ML face landmark detection finds jaw width, cheekbone span, forehead height, and chin shape from a single webcam frame.
The shopper's face is classified into one of five shapes — oval, round, square, heart, oblong — with a confidence score.
A recommendation rules engine maps the classified shape to frame silhouettes that flatter it, pulled from the brand's catalog.
Five shapes. One shortlist.
The recommendation rules engine is brand-aware — feed it your catalog's silhouette tagging and it surfaces matches from your own SKUs, not generic shapes.
Turn your catalog into a shortlist.
Embed the analyzer in Commerce Studio, or talk to us about tuning the recommendation rules to your brand's silhouette taxonomy.