In order to combat biases in facial-recognition systems, IBM plans to use 1 million photos it acquired from Flickr to train the technology, CNBC reports. IBM acquired 1 million images from Flickr in order to assist its facial-recognition systems to identify faces in a more fair and accurate manner. The pictures contain a diverse set of data subjects, including of individuals of different ages, genders and facial symmetries. "Many prominent datasets used in the field are too narrow and fall short in coverage and balance," IBM Lead Scientist John Smith said. "The data does not reflect the faces we see in the world." Meanwhile, San Francisco Supervisor Aaron Peskin proposed legislation to ban facial-recognition surveillance in the city. Editor's Note: Kate Kaye writes about the U.S. National Institute of Standards and Technology's facial-recognition accuracy test in this piece for Privacy Tech.
If you want to comment on this post, you need to login.