Computer-aided medical image segmentation has been applied widely in diagnosis and treatment to obtain clinically useful information of shapes and volumes of target organs and tissues. In the past several years, convolutional neural network (CNN) based methods (e.g., U-Net) have dominated this area, but still suffered from inadequate long-range information capturing. Hence, recent work presented computer vision Transformer variants for medical image segmentation tasks and obtained promising performances. Such Transformers model long-range dependency by computing pair-wise patch relations. However, they incur prohibitive computational costs, especially on 3D medical images (e.g., CT and MRI). In this paper, we propose a new method called Dilated Transformer, which conducts self-attention for pair-wise patch relations captured alternately in local and global scopes. Inspired by dilated convolution kernels, we conduct the global self-attention in a dilated manner, enlarging receptive fields without increasing the patches involved and thus reducing computational costs. Based on this design of Dilated Transformer, we construct a U-shaped encoder-decoder hierarchical architecture called D-Former for 3D medical image segmentation. Experiments on the Synapse and ACDC datasets show that our D-Former model, trained from scratch, outperforms various competitive CNN-based or Transformer-based segmentation models at a low computational cost without time-consuming per-training process.
翻译:在诊断和治疗中广泛应用计算机辅助医疗图象分割法,以获取关于目标器官和组织形状和数量的信息。在过去几年中,基于神经神经网络(CNN)的进化方法(例如U-Net)占据了该领域的主导地位,但仍受到长期信息采集不足的影响。因此,最近的工作提出了用于医疗图象分割任务的计算机视觉变异器,并取得了有希望的性能。这些变异器模型通过计算双向补丁关系来模拟长期依赖性能。然而,它们造成了过高的计算成本,特别是在3D医疗图(例如CT和MRI)上。在本文中,我们提出了一个称为Dilate型神经网络变异器(CNN)的新方法,该方法在地方和全球范围内对双向补配补关系进行自我关注。在变异性图层中,我们以基于模型的方式进行全球自留,扩大可接受性域域,同时不增加所涉的补缺位,从而降低计算成本。基于D-D级变换模型的设计,我们在不经过培训的D级变压结构中构建了一种称为U-C级变压结构的DForizeral-C结构,该结构,该结构是用来用于Slodicreal-C的D-redustry-deal-deal-chaduducal-deal-deal-deal-deal-deal-de Ad-deal-dection-deal-dection-deal-dection-deal-dection-dection-dection-dection-se-destr-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-demental-dection-dection-dection-deal-deal-deal-dection-deal-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-dection-deal-deal-deal-deal-deal-deal-deal-deal-deal-deal-deal-de-de-de-de-de-deal-de-