Yeah. It kind of is. If we trained as much for black faces as white faces, the algorithms would be better. Moreover, the background contrast for white faces is frequently worse (against backgrounds of white walls).
> If we trained as much for black faces as white faces
Doesn't Zoom have a large part of their employee base in China, and their US CEO is a Chinese immigrant? So perhaps you should replace "white" with "Asian"?
aardshark|5 years ago
rbecker|5 years ago
Doesn't Zoom have a large part of their employee base in China, and their US CEO is a Chinese immigrant? So perhaps you should replace "white" with "Asian"?