Skip to main content
. 2017 Jan 31;7:41545. doi: 10.1038/srep41545

Table 3. Performance comparison of schemas (5) and (6).

Schema Feature extraction method Kernel function Accuracy Sensitivity Specificity
(5) haar wavelet linear 0.8758 ± 0.0315 0.7976 ± 0.0517 0.9432 ± 0.0186
polynomial 0.8814 ± 0.0109 0.7853 ± 0.0263 0.9642 ± 0.0186
db1 wavelet linear 0.8758 ± 0.0315 0.7976 ± 0.0517 0.9432 ± 0.0186
polynomial 0.8814 ± 0.0109 0.7853 ± 0.0263 0.9642 ± 0.0186
sym4 wavelet linear 0.8679 ± 0.0288 0.7902 ± 0.0589 0.9348 ± 0.0210
polynomial 0.8611 ± 0.0104 0.7463 ± 0.0249 0.9579 ± 0.0194
(6) LBP linear 0.8635 ± 0.0248 0.8367 ± 0.0404 0.8865 ± 0.0223
polynomial 0.8827 ± 0.0189 0.8463 ± 0.0289 0.9118 ± 0.0287