Fig 1.
CNN automatic configuration process in the proposed method.
Fig 2.
The structure of each dense layer in the proposed basic CNN pattern.
Fig 3.
The basic CNN pattern used in the proposed method for NBIS.
Fig 4.
LA structure used in the proposed method.
Fig 5.
DICE score values resulting from segmentation by the proposed method for test samples.
Fig 6.
ASD values of the proposed method for test samples of the database.
Fig 7.
Segmentation results of one of the database images in different dimensions and ground-truth state.
Fig 8.
Best DICE score obtained for CNN model during different configuration iterations by LAs.
Table 1.
Effectiveness comparison of the proposed method in segmentation of iSeg17 test samples with previous methods based on DICE score.
Table 2.
Effectiveness comparison of the proposed method in segmentation of iSeg17 test samples with the previous method based on ASD values.