I think face recognition software was being calibrated using the phases of "social distancing" alternating with phases of regular crowds for iterations of software testing, and masks were for fine-tuning the face recognition software's skill in recognizing upper face segments, plus, extrapolating lower face segment based on upper segment and blurred lower segment.
Isn't this so face recognition can work properly?
I think face recognition software was being calibrated using the phases of "social distancing" alternating with phases of regular crowds for iterations of software testing, and masks were for fine-tuning the face recognition software's skill in recognizing upper face segments, plus, extrapolating lower face segment based on upper segment and blurred lower segment.
Always the experimental lab animalтАж.