A Real-Time Embedded System for Driver Drowsiness Detection Based on Visual Analysis of the Eyes and Mouth Using Convolutional Neural Network and Mouth Aspect Ratio

. 2024 Sep 27;24(19):6261. doi: 10.3390/s24196261

Algorithm 1 ROI Correction

1:
Input: Points: $[63, 117, 293, 346, 107, 336]$
2:
Output: Corrected_ROI
3:
$x_{a}, y_{a} \leftarrow P_{336} [x], P_{336} [y]$
4:
$x_{b}, y_{b} \leftarrow P_{63} [x], P_{63} [y]$
5:
$d_{1} \leftarrow int (distance (P_{63}, P_{117}))$
6:
$d_{2} \leftarrow int (distance (P_{63}, P_{107}))$
7:
$d_{3} \leftarrow int (distance (P_{336}, P_{293}))$
8:
$d_{4} \leftarrow int (distance (P_{293}, P_{346}))$
9:
if $x_{a} > x_{b}$ then
10:
$s t a r t_x, e n d_x \leftarrow x_{b}, (x_{a} + d_{3})$
11:
else
12:
$s t a r t_x, e n d_x \leftarrow x_{a}, (x_{b} + d_{2})$
13:
end if
14:
if $y_{a} > y_{b}$ then
15:
$s t a r t_y, e n d_y \leftarrow y_{b}, (y_{a} + d_{4})$
16:
else
17:
$s t a r t_y, e n d_y \leftarrow y_{a}, (y_{b} + d_{1})$
18:
end if
19:
if $(e n d_x - s t a r t_x) > 10 and (e n d_y - s t a r t_y) < 400$ then
20:
$s t a r t_x, s t a r t_y \leftarrow s t a r t_x - 10, s t a r t_y - 10$
21:
$e n d_x, e n d_y \leftarrow e n d_x + 10, e n d_y + 10$
22:
Corrected_ROI ← [start_y:end_y, start_x:end_x]
23:
end if