Research Article
Is Intensity Inhomogeneity Correction Useful for Classification of Breast Cancer in Sonograms Using Deep Neural Network?
Table 1
Classification results of the private database (the displayed value: original/corrected).
| ā | SDAE | AlexNet | Inception v3 | ResNet | DenseNet |
| TP | 8/11 | 5/6 | 6/7 | 7/7 | 6/6 | TN | 9/11 | 7/4 | 6/5 | 6/7 | 8/7 | FP | 4/2 | 7/10 | 8/9 | 8/7 | 6/7 | FN | 6/3 | 8/7 | 7/6 | 6/6 | 7/7 | Precision | 0.67/0.85 | 0.42/0.38 | 0.43/0.44 | 0.5/0.5 | 0.5/0.46 | Recall | 0.57/0.79 | 0.38/0.46 | 0.46/0.54 | 0.46/0.54 | 0.46/0.46 | Specificity | 0.69/0.85 | 0.5/0.28 | 0.43/0.36 | 0.57/0.5 | 0.57/0.5 | Accuracy | 0.63/0.82 | 0.44/0.37 | 0.44/0.44 | 0.51/0.52 | 0.52/0.48 | F-measure | 0.62/0.85 | 0.4/0.41 | 0.44/0.48 | 0.48/0.52 | 0.48/0.46 |
|
|