Research Article
Conditional Random Fields and Supervised Learning in Automated Skin Lesion Diagnosis
Table 1
Comparison of our PP model’s ability to segment lesions to our CRF model and 5 previously published methods.
| | | Performance | | | Method | PP (nearest pt.) | Method | | Sens | Spec | Sens | Spec |
| CRF | 116 | 0.845 | 0.924 | 0.843 | 0.921 | KPP [6] | 116 | 0.765 | 0.770 | 0.941 | 0.763 | JSEG [7] | 91 | 0.627 | 0.987 | 0.677 | 0.980 | DTEA [8] | 116 | 0.597 | 0.986 | 0.638 | 0.985 | SRM [9] | 112 | 0.790 | 0.946 | 0.773 | 0.957 | FSN [10] | 116 | 0.808 | 0.934 | 0.814 | 0.939 |
|
|