A Practical Cosmetic Try-on System using OpenCV

Sohee Yoon

犀利士 “has-text-align-center”>Korea International School

Abstract

There are a growing number of applications that allow customers to virtually apply cosmetic items in order to try out various makeup styles and sample new products before making an online purchase. The integration of facial recognition algorithms with augmented reality technologies facilitates this. This method has limitations due to the fact that the sampled color is shown on top of the user’s skin tone without taking into account the combination of two colors and the transparency of the mixed color. Thus, the following research paper provides an alternative method for producing precise color combinations. The system comprises two primary components: First, perform gamma correction and adjust the image’s brightness and contrast. The color of the user will then be extracted. Second, once the user selects the desired product color, the system mixes the user’s skin tone with the color of the product. The evaluation section demonstrated the practicability of the proposed system. At the conclusion of the study, I make a heuristic proposal for a user- satisfaction-improving recommendation system based on object detection and the k-mean clustering algorithm in machine learning.

Please click to read the paper.