Kozmetik Yüz Algılama Uygulaması Karşılaştırmalı Kullanılabilirlik Değerlendirmesi

As cosmetic facial detection applications become increasingly popular people start to ask what ethics and usability are at play. This research explores the usability and ethical challenges of cosmetic facial detection applications under varying cosmetic conditions such as makeup and lighting. A mixed methods approach of quantitative data from 64 participants and qualitative insights from five in depth interviews identify key challenges, including low detection accuracy, interface inefficiencies, and high privacy concerns. With heavy makeup or in bad lighting, users said they found the apps were inconsistent, and they chafed at what they considered biases toward lighter skin and simpler looks. However, privacy policies were also often considered too complex and inaccessible to further reduce user trust. Features such as virtual try-ons and skin analysis were well received but slow performance and fuzzy privacy were mourned by users. This suggests that algorithmic accuracy should be improved, biases should be addressed, and privacy policies made simpler as well as user education provided. The goals of these steps are to build trust, increase diversity, and enhance user experience in the hopes that these technologies will not only be accepted, but also exceeded by these diverse consumer bases of the beauty industry.