Alternative Non-Destructive Approach for Estimating Morphometric Measurements of Chicken Eggs from Tomographic Images with Computer Vision

. 2024 Dec 14;13(24):4039. doi: 10.3390/foods13244039

Algorithm 2 Training Stage 2

Require:
$D$ ′, $m o d e l_{θ}$ ▹ Dataset
1:
$m o d e l_{θ}$ .require_grad(False)
2:
$m o d e l_{θ}$ .attach_head(require_grad=True)
3:
for epoch in ${0, 1, \dots, E}$ do
4:
for all batch in $D$ do
5:
$e g g s \leftarrow b a t c h [0]$ ▹ 3D CT egg image
6:
$m e a s u r e s \leftarrow b a t c h [1]$ ▹ Morphological measures
7:
$p r e d i c t e d_m e a s u r e s \leftarrow m o d e l_{θ} (e g g s)$ ▹ Makes a prediction
8:
$l o s s \leftarrow loss (m e a s u r e s, p r e d i c t e d_m e a s u r e s)$
9:
model.backprop(loss)
10:
end for
11:
end for
12:
Output $m o d e l_{θ}$