Makeup is an entertainment that allows people to present themselves in a way that is close to their ideal. People apply foundation makeup evenly to cover skin imperfections, but it is not easy to distinguish between areas where foundation is applied or not. Therefore, problems like applying too much foundation or forgetting to apply foundation on some areas are likely to occur. Our project aims to realize a makeup support system that can visualize the state of foundation application in real-time to eliminate uneven application. As an initial step, we conducted a large-scale crowdsourcing survey on makeup, precisely foundation, and investigated a method that uses machine learning to classify between images of skin with foundation applied and of bare skin. As a result, we found that we could distinguish between both with high accuracy of 82.3%.