Using Convolutional Neural Network to Improve Decision Making in Cosmetic Purchases

In this era, of course everyone knows all about style and beauty. One of them is lipstick. You can even see it in a mall/beauty shop if there's no lipstick, that's quite a strange thing. In every single place (mall/beauty shop), of course, they have their own marketing techniques from discounts, giveaways and others. The concept of a virtual lipstick tester is applying a lipstick color to the customer's lips on a digital screen according to the user's choices. This experiment uses CNN which has the advantage of giving more accurate color prediction. We proposed CNN color prediction algorithm to give color prediction by considering user base lip color to enhance color reproduction by giving more accuracy. Through this research we find that the CNN color prediction algorithm is capable of improving color reproduction significantly with 86.59% accuracy result.