UNet-2022: Exploring Dynamics in Non-isomorphic Architecture
This work addresses a key bottleneck in medical image segmentation by providing a clearer design rationale, though it is incremental as it builds on existing hybrid architectures.
The authors tackled the lack of intuitive explanation for hybrid self-attention and convolution models in medical image segmentation by proposing UNet-2022, a parallel non-isomorphic block that outperforms baselines by up to 4% and surpasses nnUNet in multiple tasks.
Recent medical image segmentation models are mostly hybrid, which integrate self-attention and convolution layers into the non-isomorphic architecture. However, one potential drawback of these approaches is that they failed to provide an intuitive explanation of why this hybrid combination manner is beneficial, making it difficult for subsequent work to make improvements on top of them. To address this issue, we first analyze the differences between the weight allocation mechanisms of the self-attention and convolution. Based on this analysis, we propose to construct a parallel non-isomorphic block that takes the advantages of self-attention and convolution with simple parallelization. We name the resulting U-shape segmentation model as UNet-2022. In experiments, UNet-2022 obviously outperforms its counterparts in a range segmentation tasks, including abdominal multi-organ segmentation, automatic cardiac diagnosis, neural structures segmentation, and skin lesion segmentation, sometimes surpassing the best performing baseline by 4%. Specifically, UNet-2022 surpasses nnUNet, the most recognized segmentation model at present, by large margins. These phenomena indicate the potential of UNet-2022 to become the model of choice for medical image segmentation.