Advanced Makeup Try-On Solution: AI, AR, and Big Data Integration

AI-driven AR makeup try-on with precise shade detection and customization


Created by:

Key takeaways

This is some text inside of a div block.

The project aimed to develop a custom makeup try-on solution leveraging advanced technologies like Big Data, Artificial Intelligence (AI), Computer Vision (CV), and web-based Augmented Reality (AR) for real-time face augmentation. It addressed the limitations of the existing Perfect Corp software by enhancing the capability to try multiple makeup combinations and create unique looks.

Utilizing web-based AR technology, the solution enables users to experiment with entire makeup looks across the full Pantone scale and multiple products simultaneously. AI and Big Data integration introduce a unique feature recommending products based on user demographics, location, complexion, face shape, preferences, and past purchases. A key innovation allows users to upload images of celebrity or personal makeup styles, analyzed against a palette of over 10,000 shades to recommend matching products accurately.

The implementation employed adaptive facial landmark technology for precise facial feature and complexion identification, facilitating real-time try-ons. Integration with Shopify enables seamless purchase of individual products or complete sets used in the uploaded image look creation.

Key components of the solution include:

1. Custom Framework Solution: Developed to assign multiple physical product SKUs to one Perfect Corp SKU, expanding the product and shade portfolio.

2. Big Data Utilization: Collected from user try-on experiences to enhance recommendation accuracy.

3. Complexion Detection: Enhanced through data cleaning, shadow and light removal via color frequency analysis, continuous machine learning, and data mining.

4. Web-Based AR: Compatible with desktop and mobile devices for widespread accessibility.

5. Portfolio Management Tool: Allows easy navigation of SKU lists, management of available palettes, and real-time try-ons.

6. Real-Time Facial Feature Detection: Employs neural network systems for accurate analysis.

7. Color Detection and Recognition: Uses the CIEDE2000 color distance formula and K-mean based color clustering for precise color matching.

Overall, the project revolutionizes virtual makeup try-on experiences by integrating cutting-edge technologies to offer a comprehensive, personalized, and user-friendly platform. It aims to provide a seamless blend of creativity, functionality, and shopping convenience, enhancing user engagement and satisfaction in the beauty industry.

Rrahul Sethi
June 25, 2024
5 min read