Removal of False Blood Vessels Using Shape Based Features and Image Inpainting

Table 2

Performance measures comparison of different features and feature selection techniques.

Data	Accuracy (%)	Sensitivity (%)	Specificity (%)
	mean ± std	mean ± std	mean ± std
Full data	87.35 ± 0.0878	94.69 ± 0.1282	86.19 ± 0.0623
Pca2 (top2)	82.32 ± 0.0055	98.60 ± 0.0053	24.88 ± 0.0133
Pca3 (top3)	88.00 ± 0.0071	95.25 ± 0.0045	67.80 ± 0.0192
Pca4 (top4)	86.47 ± 0.0673	96.38 ± 0.0181	72.24 ± 0.0628
Pca5 (top5)	87.33 ± 0.1244	91.85 ± 0.0602	90.95 ± 0.0602
Pca6 (top6)	86.69 ± 0.0692	91.90 ± 0.1018	91.31 ± 0.0569
Wil2 (top2)	87.78 ± 0.0032	95.00 ± 0.0037	64.03 ± 0.0101
Wil3 (top3)	88.00 ± 0.0055	94.43 ± 0.0058	77.20 ± 0.0255
Wil4 (top4)	87.56 ± 0.1673	93.36 ± 0.2381	74.36 ± 0.0315
Wil5 (top5)	87.56 ± 0.1482	94.88 ± 0.2151	93.52 ± 0.0990
Wil6 (top6)	88.20 ± 0.1531	94.98 ± 0.2208	91.18 ± 0.0939