This AR-based cosmetic lens try-on app offers users a personalized and interactive shopping experience, similar to leading beauty brands like L'Oréal and Sephora. Designed for mobile devices, the app allows users to virtually try on cosmetic contact lenses in real time, using augmented reality. By analyzing the user's skin tone through facial landmark detection and skin tone classification (based on 5–7 common Asian skin tones), the app recommends 3–4 lens shades that best complement their complexion. Eye region tracking ensures accurate overlay of virtual lenses on the user's eyes, creating a realistic and immersive try-on experience. The app features a friendly, easy-to-use interface where users can explore recommended lenses or manually browse the collection. Advanced computer vision techniques, including dlib and OpenCV, are used for face detection, eye tracking, and skin tone analysis, while ARCore (for Android) handles the AR rendering. A dedicated lens dataset with categorized tones and styles powers the recommendation engine. This solution enhances online cosmetic shopping by reducing guesswork, increasing confidence in purchases, and providing a fun and engaging user experience. It bridges the gap between digital convenience and physical product testing, making it ideal for beauty retailers targeting tech-savvy customers.
Tools: ML,Unity,Android Studio,Python,Kotlin,Flask,Figma,OpenCv
Department: Department of Business Studies
Project Poster
