Table 3 Performance of DenseNet model
From: Application of entire dental panorama image data in artificial intelligence model for age estimation
DenseNet | ||||||
|---|---|---|---|---|---|---|
Number of images | train:13,220, val:1653, test:1653 | |||||
parameters | batch8, epoch40, resize256 | |||||
performance | Loss | Acc | Precision | Recall | F1-score | |
Type 1 grouping | Basic prediction | 0.5899 | 0.1016 | 0.0579 | 0.0583 | 0.058 |
with ± 3 years deviation | 0.5905 | 0.2813 | 0.1776 | 0.1768 | 0.1764 | |
Type 2 grouping (heuristics) | Basic prediction | 0.412 | 0.3146 | 0.2115 | 0.2117 | 0.2072 |
with ± 3 years deviation | 0.4116 | 0.7641 | 0.6632 | 0.6658 | 0.6583 | |