error correctionAblation experiments are carried out on dataset #1 and dataset #2 to verify the effectiveness of the proposed network This section verifies the effectiveness of the backbone net-work o

Ablation experiments were carried out on dataset #1 and dataset #2 to verify the effectiveness of the proposed network. This section examines the effectiveness of the CNN backbone network in the encoder and the effectiveness of multi-scale feature fusion in the decoder.

1. Effectiveness of CNN backbone The main component of the encoder in the multi-objective segmentation network proposed in this paper is the CNN backbone network. To select a more suitable backbone network, we evaluated various backbone networks on dataset #1 and dataset #2. In these experiments, all parameters, except for the backbone network, remained consistent. We used Resnet-50, Densenet, CoTNeXt-50, and CoT-Net-50 as the backbone networks and trained and tested them on the constructed datasets. The test results are presented in Table V and Table VI. It can be observed that CoTNet-50 achieved better results