A Convolutional Neural Network for Lentigo Diagnosis
97
Healthy (PS
≥ 0.5)
PS=0.997
PS=0.613
Lentigo (PS
< 0.5)
PS=0.008
PS=0.308
Fig. 6. Correct classification examples of RCM images for Healthy and Lentigo patients
classified by the proposed method.
Fig. 7. The only false classification (PS = 0.0005).
Table 4. Comparison performance with state of the art methods.
Lentigo detection method Accuracy
Halimi et al. 2017 [10]
84.4%
Halimi et al. 2017 [11]
97.7%
Proposed method
98,14%
97
Healthy (PS
≥ 0.5)
PS=0.997
PS=0.613
Lentigo (PS
< 0.5)
PS=0.008
PS=0.308
Fig. 6. Correct classification examples of RCM images for Healthy and Lentigo patients
classified by the proposed method.
Fig. 7. The only false classification (PS = 0.0005).
Table 4. Comparison performance with state of the art methods.
Lentigo detection method Accuracy
Halimi et al. 2017 [10]
84.4%
Halimi et al. 2017 [11]
97.7%
Proposed method
98,14%