Table 2.

Effect of preprocessing on Dice loss of the training dataset

Sample	Without preprocessing	With preprocessing	Improvement (%)
S0033	0.154	0.142	7.7
S0066	0.150	0.126	15.8
S0150	0.123	0.112	8.8
S0300	0.114	0.107	5.8
S0600	0.115	0.100	13.2
S3600	0.110	0.099	10.1

Assuming that the value of “Without processing” is A and the value of “With processing” is B, “Improvement” was calculated by (A–B)/A