in this paper我们提出crossmodel去对患侧病灶进行更精准的分割。我们设计了一种基于UNet和Transformer的网络结构去学习病灶分割。如图显示我们的crossmodel通过准确地互补双视角的肿块的特征信息实现更精准合理的肿块分割。翻译成英文

In this paper, we propose a crossmodel approach for more accurate segmentation of lesions on the affected side. We designed a network architecture based on UNet and Transformer to learn lesion segmentation. As shown in the figure, our crossmodel achieves more precise and reasonable lesion segmentation by accurately complementing the feature information of the tumors from both perspectives.