Pinterest has launched new changes that allow users to filter beauty and beauty search results by ranges of skin tones, as a way to make it easier to find what they are looking for. The changes occur after the social network platform conducted a survey and found that 70 percent of people used the service to find and save looks and everyday styles.
A medium publication of the Pinterest engineering account says that the process of programming search filters to detect skin tone in images was difficult, since lighting, shadows and other factors were all variable to consider.
To solve the problem, Pinterest used the machine learning of a ModiFace third-party beauty application that already had a facial data library suitable for artificial intelligence use. Pinterest then trained the algorithms not to read white skin in dark shadows as dark skin tone, and corrected other errors by repeated tests.
Pinterest also noted that when users searched for skin tones, that information will not be stored and will not affect the advertisements posted to users. The property is launching today in beta, says the company.
"Our future work will generally focus on improving the accuracy of the results and bringing the experience to more platforms, we will try new methods of rewriting and combining queries into actionable content and we hope to improve our search ranking models to have better in account for the selected skin tone, "writes Laksh Bhasin, a Pinterest engineer who works on his quality search team. "Finally, skin tones are just the beginning of building a more inclusive search, and we hope to help Pinners find more personalized results by offering more ways to narrow their search."