TransFusion: Multi-view Divergent Fusion for Medical Image Segmentation with Transformers

Combining information from multi-view images is crucial to improve the performance and robustness of automated methods for disease diagnosis. However, due non-alignment characteristics images, building correlation data fusion across views largely remain an open problem. In this study, we present TransFusion, a Transformer-based architecture merge divergent imaging using convolutional layers powerful attention mechanisms. particular, Divergent Fusion Attention (DiFA) module proposed rich cross-view context modeling semantic dependency mining, addressing critical issue capturing long-range correlations between unaligned different image views. We further propose Multi-Scale (MSA) collect global correspondence multi-scale feature representations. evaluate TransFusion on Multi-Disease, Multi-View & Multi-Center Right Ventricular Segmentation in Cardiac MRI (M &Ms-2) challenge cohort. demonstrates leading against state-of-the-art opens up new perspectives integration towards robust medical segmentation.