Luxury cosmetics

Face analysis for product recommendations

A luxury cosmetics player wanted to offer its customers a personalized skin diagnosis directly in store: a photo, a score, identified criteria and a tailor-made product recommendation. We designed the Deep Learning models that make this experience possible, which are now deployed in stores and have resulted in a scientific publication.

Problem

The recommendation of cosmetic products in stores was based on the subjective expertise of the consultants. No tool made it possible to objectify the condition of a client's skin, to identify the determining dermatological criteria and to deduce a personalized product recommendation. The customer wanted to transform the in-store experience with an application that could score a face from a simple photo, visualize the criteria in an understandable way, and generate images showing the impact of the changed attributes all in real time, in front of the customer.

Vue rapprochée d’une coupe transversale colorée d’une géode montrant des couches concentriques de minéraux en jaune, marron, rouge et vert.

Solution

What we built

We deployed 2 Data Scientists and 1 Data Engineer to design a complete skin analysis system by Computer Vision, from the model to the production in store.

Step 1 — Architecture choice and preprocessing. Exploration and selection of model architectures best suited to the problem, with adaptation of state-of-the-art architectures. Definition of preprocessing pipelines to guarantee robustness on photos taken in real conditions (shop lighting, skin diversity).

Step 2 — Training and optimization. Optimization of hyperparameters, parallelization of training and establishment of an internal framework to accelerate iterations. The model scores the face on several dermatological criteria from a single photo.

Step 3 — Interpretability and image generation. Development of a module for visualizing internal features to identify the determining criteria for scoring. Generation of modified images to show the impact of each attribute in an intuitive and visual way for the client.

Step 4 — Production started in store. Deployment of the application in store, complete documentation and transfer to internal teams. The system works in real time in front of the client.

Projects in the same category

See all projects

Supply Chain Optimization Application

A pharmaceutical distribution player had to rethink its entire supply chain: pharmacy assortment, inventory management and delivery channels. Operational research alone was no longer enough. We built the application that transformed its operations.

Social media trend detection

In a market where consumer behaviors change faster than decision cycles, a customer needed to anticipate trends instead of experiencing them. We built the platform that turns social media noise into actionable signals.

Yield of agricultural fields

The climate subsidiary of a French insurance leader needed to predict field yields across Germany to price a new drought insurance offer. Internal data was not enough. We built the predictive models that made the product marketable.

Customer needs analysis platform

A large French group needed to understand the current and future needs of its customers by aggregating a massive volume of consumer data. After a POC of 8 weeks, we industrialized a complete platform deployed on all French entities and is now being extended at the group level.

Skin scoring and analysis

A major player in luxury cosmetics had developed a skin scoring algorithm. Problem: no one could verify what the AI was based on to make its diagnosis. We built the visualization system that makes predictions transparent and deployable in stores and on mobile.

Beaconing detection

A large group's security team needed to identify beaconing signals, regular network traffic sent by potentially compromised machines to servers controlled by attackers. The volume of logs made manual detection impossible. We built the anomaly detection system capable of processing this data on a large scale.