Transform and multi-scale feature fusion-based medical image segmentation method and system

The invention discloses a medical image segmentation method and system based on Transform and multi-scale feature fusion, and relates to the technical field of image processing. Richer multi-scale information is extracted through a multi-scale block formed by stacked convolution blocks during coding; meanwhile, a scale perception pyramid fusion module is further used for dynamically fusing multi-scale context information in high-level features, and a cross fusion Transform module is used on jump connection to replace simple jump connection to solve the problem of semantic feature inconsistency, so that up-sampling information is better recovered; during up-sampling of a decoder, an attention mechanism is used for highlighting important features and inhibiting irrelevant features to help up-sampling to better recover feature information, and obviously better segmentation performance can be obtained. 基于Transformer和多尺度特征融合的医学图像分割方法及系统，涉及图像处理技术领域，本发明在编码时通过堆叠的卷积块形成的多尺度块提取更加丰富的多尺度信息，同时进一步使用了一个尺度感知的金字塔融合模块来动态融合高层特征中