Table 4 Metrics and variances of different methods on the bone tumor dataset.

From: Bone tumor recognition strategy based on object region and context representation in medical decision-making system

Model

mAccos

mPreos

mReos

mF1os

mIOUos

mDSCos

FCN-8 s

0.9756 ± 0.04

0.9353 ± 0.07

0.8827 ± 0.05

0.9068 ± 0.04

0.8245 ± 0.045

0.8593 ± 0.04

PSPNet

0.9723 ± 0.05

0.8563 ± 0.09

0.8882 ± 0.05

0.8725 ± 0.03

0.7723 ± 0.04

0.8705 ± 0.06

FPN

0.9795 ± 0.09

0.9014 ± 0.11

0.91940.07 ± 

0.9103 ± 0.06

0.8516 ± 0.05

0.8788 ± 0.08

Unet

0.9887 ± 0.08

0.9122 ± 0.09

0.9124 ± 0.07

0.9123 ± 0.06

0.8672 ± 0.04

0.8927 ± 0.05

MSFCN

0.9817 ± 0.07

0.8801 ± 0.06

0.8936 ± 0.04

0.8868 ± 0.05

0.8439 ± 0.02

0.8741 ± 0.03

MSRN

0.9780 ± 0.04

0.8393 ± 0.03

0.8945 ± 0.05

0.8660 ± 0.04

0.8134 ± 0.05

0.8787 ± 0.03

DCHS

0.6533 ± 0.04

0.7642 ± 0.06

0.7031 ± 0.06

0.7324 ± 0.05

0.7257 ± 0.05

0.7423 ± 0.04

MTCF

0.8704 ± 0.05

0.8893 ± 0.07

0.8745 ± 0.06

0.8818 ± 0.05

0.8036 ± 0.04

0.8141 ± 0..05

Our(IMEocr)

0.9756 ± 0.03

0.9029 ± 0.05

0.9102 ± 0.04

0.9065 ± 0.07

0.8694 ± 0.04

0.8968 ± 0.04

Our(Pre + IMEocr)

0.9899 ± 0.03

0.9273 ± 0.04

0.9325 ± 0.04

0.9249 ± 0.05

0.8694 ± 0.03

0.8968 ± 0.03

Our(Pre + IMEocr + SF)

0.9943 ± 0.02

0.9349 ± 0.03

0.9426 ± 0.03

0.9387 ± 0.05

0.8874 ± 0.03

0.9185 ± 0.03

Back to article page