Table 4 Accuracy of image classification of ResNet-101 model for each tooth position
From: Fully automated film mounting in dental radiography: a deep learning model
Tooth position | Accuracy | ||
|---|---|---|---|
Internal testing Taipei CGMH | External testing | ||
Linkou | Taoyuan | ||
16 ~ 18 | 0.954 | 0.946 | 0.961 |
16 ~ 14 | 0.942 | 0.932 | 0.941 |
13 ~ 15 | 0.950 | 0.934 | 0.945 |
12 ~ 22 | 0.972 | 0.952 | 0.962 |
23 ~ 25 | 0.962 | 0.958 | 0.956 |
24 ~ 26 | 0.984 | 0.975 | 0.971 |
26 ~ 28 | 1.000 | 0.986 | 0.975 |
46 ~ 48 | 0.992 | 1.000 | 0.986 |
46 ~ 44 | 0.946 | 0.938 | 0.952 |
43 ~ 45 | 0.979 | 0.962 | 0.963 |
42 ~ 32 | 0.983 | 0.952 | 0.964 |
33 ~ 35 | 0.963 | 0.952 | 0.973 |
34 ~ 36 | 0.972 | 0.983 | 0.971 |
36 ~ 38 | 0.976 | 0.962 | 0.973 |
Occlusal Upper | 1.000 | 1.000 | 1.000 |
Occlusal Lower | 1.000 | 1.000 | 1.000 |
53 ~ 16 | 0.946 | 0.933 | 0.958 |
52 ~ 62 | 0.992 | 0.982 | 0.985 |
63 ~ 26 | 0.968 | 0.963 | 0.965 |
83 ~ 46 | 0.984 | 0.956 | 0.978 |
72 ~ 82 | 1.000 | 0.976 | 0.985 |
73 ~ 36 | 0.978 | 0.964 | 0.963 |
BW-Right | 0.992 | 0.994 | 0.986 |
BW-Left | 0.975 | 0.986 | 0.971 |
VBW-Right anterior | 0.967 | 0.952 | 0.946 |
VBW-Right posterior | 0.976 | 0.963 | 0.966 |
VBW-Left anterior | 0.942 | 0.952 | 0.943 |
VBW-Left posterior | 0.934 | 0.921 | 0.938 |
Panorex | 1.000 | 1.000 | 1.000 |
TMJ | 1.000 | 1.000 | 1.000 |
Cephalometric posterior-anterior | 1.000 | 1.000 | 1.000 |
Cephalometric Lateral | 1.000 | 1.000 | 1.000 |