Microsoft has revealed new facial recognition software that’s more capable of identifying men and women with darker skin.
The company’s Face API, part of Azure Cognitive Services, is able to better identify those with darker skin, in comparison to previous iterations of the technology.
Microsoft says its system’s recognition reduces the error rates for darker-skinned men and women by up to 20 times. Further, it was able to reduce the error rates for all women by nine times.
Microsoft says the higher fail rates on females with darker skins is industrywide.
“If a facial recognition system is to perform well across all people, the training dataset needs to represent a diversity of skin tones as well as factors such as hairstyle, jewelry and eyewear,” said Microsoft.
Microsoft was able to improve the artificial intelligence by focusing on blind spots in training data. Ece Kamar, a senior researcher for Microsoft, spotted an under representation of darker skinned women in the training data, which may have led to the AI systems error rates.
“Collecting more data that captures the diversity of our world and being careful about how to measure performance are important steps toward mitigating these issues,” she said.
As a darker skin man myself, I have found some software to be better than others. While I’ve never used Apple’s Face ID, I have found many faults with Samsung’s iris scanning and facial recognition software. Meanwhile, Huawei’s facial recognition software works great under nearly all circumstances.