Extraction of Prostatic Lumina and Automated Recognition for Prostatic Calculus Image Using PCA-SVM

Table 1

Each column shows results of one time of 5-fold cross validation, which includes the five steps described.

	1	2	3	4	5	6	7	8	9	10
ATT (s)	0.1438	0.144	0.1532	0.1436	0.1314	0.1312	0.1374	0.1594	0.153	0.1346
AART (%)	100	100	100	100	100	100	100	100	100	100
ATA (%)	92.8	92	93.6	92.8	94.4	93.6	94.4	91.2	94.4	92
TNS	91	90	93	93	93	92	94	92	95	90
FPS	6	7	4	5	5	5	4	5	2	7
FNS	3	3	4	4	2	3	3	6	5	3
TPS	25	25	24	23	25	25	24	22	23	25
Sensitivity (%)	89.5	89.3	86.7	83.8	92.5	89.8	91.3	78.6	84.5	91.4
Specificity (%)	94	92.8	95.9	94.7	94.9	94.8	96.3	94.7	97.8	92.3

ATT: average training time; AART: average accuracy rate of training; ATA: average test accuracy; TNS: true negative samples; FPS: false positive samples; FNS: false negative samples; TPS: true positive samples.