The system achieved 91% accuracy, outperforming human evaluations, and enabled the automated selection of the most suitable cosmetics for various skin conditions.
Meet our client
Client:
Industry:
Market:
Technology:
Client’s Challenge
A global cosmetics company sought to develop a system capable of assessing skin conditions based on evenness, hydration, shininess, and age. The goal was to create a solution that could provide personalized skincare recommendations, automating a process traditionally performed by dermatologists.
Our Solution
We built a deep neural network that accurately analyzes different skin features. The project involved curating two datasets – one professionally assessed by dermatologists and another crowdsourced – to create a well-calibrated system. Extensive data preprocessing, augmentation, and testing of DNN architectures ensured accuracy.
Client’s Benefits
The system achieved 91% accuracy, outperforming human evaluations, and enabled the automated selection of the most suitable cosmetics for various skin conditions. This streamlined the cosmetics selection process, making it more personalized.
