๐ค AI model boosts medical image classification accuracy
A new medical image classification algorithm was developed using a hierarchical and complementary attention-enhanced Swin Transformer model. Researchers from Hebei University of Architecture in China introduced this framework to handle challenges in lesion scale and tissue complexity. The study tested it on eight subsets of MedMNIST v2 with a 224x224 resolution, including BreastMNIST and PneumoniaMNIST. This architecture uses specialized blocks to improve multi-scale feature representation and contextual modeling effectively. The proposed model shows strong robustness across various medical imaging modalities for diagnosis support. ๐ค