Autonomous Robots for Deep Mask-Wearing Detection in Educational Settings during Pandemics

Table 7

Differences between the base and improved models.

Differences	Base model	Improved model
Number of models	1 model	3 models
Categories	Correct, incorrect, and without	Correct, incorrect, and without
Dataset size	10,000 images	12,000 images
Mask colors	Bright mask colors only	Bright and dark mask colors
Face angles	Facing the camera directly	Face angles between 45 and 90 degrees
Frame rate	12 FPS	10 FPS
Inference time	83 ms	100 ms
Avg. precision	0.775	0.914
Avg. recall	0.654	0.654