Skip to main content

Table 9 Comparison with state-of-the-art methods

From: A multilevel features selection framework for skin lesion classification

RefYearDatasetMethodOA (%)
[65]2016\(PH^{2}\)ABCD rule90.00
[66]2016\(PH^{2}\)wavelet transform with morphological operations93.87
[15]2017\(PH^{2}\)multistage fully convolutional network94.24
[67]2017\(PH^{2}\)color and texture features96.00
[68]2018ISBI-2017regularised discriminant learning83.20
[13]2018ISBI-2017fully convolutional residual networks & lesion index calculation unit85.70
[69]2018ISBI-2017Ensemble Of Deep Neural Networks84.8%
[18]2018ISIC-MSKprobabilistic distribution and best features selection97.20
Proposed2019ISBI-2017ECNCA95.90
Proposed2019ISIC-UDAECNCA97.10
Proposed2019ISIC-MSKECNCA99.20
Proposed2019\(PH^{2}\)ECNCA98.80