Scenes	Supported algorithms	Description
	Algorithm accuracy	Face recognition (1: 1): 99%@0.01% (accuracy rate is 99% under 0.01% false rate) 95%@0.0001% (95% accuracy under 0.0001% false positive rate)
	Angle requirements	Yaw ± 30 degrees, roll ± 30 degrees, pitch ± 15 degrees
	Face requirements	Face size is more than 80 * 80 pixels in images
Face live detection	Angle requirements	Yaw ± 30 degrees, roll ± 30 degrees, pitch ± 15 degrees
	Face requirements	Face size is more than 80 * 80 pixels in images
	Algorithm accuracy	98%@0.01% (real person pass rate @false person pass rate)
	Non-living category	Non-living detection for pictures, videos, coated paper, face fake models, etc.
Face detection and tracking	Snapshots	More than 30 people / frame
	Snapshot frame rate	25fps
	Smallest face	30*30
	Adaptation angle	Yaw ± 45 degrees, roll ± 45 degrees, pitch ± 30 degrees
	Attribute content	Gender, age, glasses, expression (angry, calm, happy, sad,

Face attribute recognition

surprised), mask, race, beard, race, hat detection, hard hat

detection

Algorithm accuracy

Age accuracy (± 5 years): more than 85%; Gender accuracy: 96% or more;

Skin tone accuracy: above 97%; Beard accuracy: above 97%; Glasses accuracy: more than 98%; Mask accuracy: more than 98%; Hat accuracy: above 97%;

Expression accuracy: 88% or more

Product Tags:

human body temperature sensor

human temperature sensor

Inquiry Cart 0