Brain tumor image segmentation based on improved FPN

BMC Medical Imaging

Table 3 Performance comparison of the proposed model with state-of-art methods

Literature	Dataset/Image Size	DSC	JAC/IoU index	Accuracy
(Dong N et al.,2017) [23]	10infants	85.03	-	-
(Pereira S et al.,2017) [24]	65subjects	88	-	-
(Zhao X et al.,2017) [25]	449patients	84	-	-
(Akkus Z et al.,2017) [26]	159images	-	-	87.7
(Hashemzehi R et al.,2017) [27]	TCGA	84		95
Our improve the FPN method	TCGA	92	86	99.1

ISSN: 1471-2342