The paper "Medical Transformer: Gated Axial-Attention for Medical Image Segmentation" explores the use of transformer-based architectures for medical image segmentation, addressing the limitations of convolutional neural networks (CNNs) in capturing long-range dependencies. The authors propose a gated axial-attention model that introduces an additional control mechanism in the self-attention module, enhancing the ability to learn global and local features. They also introduce a Local-Global (LoGo) training strategy, which operates on both the whole image and patches to improve performance. The proposed Medical Transformer (MedT) is evaluated on three medical image segmentation datasets and shown to outperform both CNNs and other transformer-based architectures. The key contributions include a gated position-sensitive axial attention mechanism, effective LoGo training methodology, and the MedT model, which demonstrates improved performance in medical image segmentation tasks.The paper "Medical Transformer: Gated Axial-Attention for Medical Image Segmentation" explores the use of transformer-based architectures for medical image segmentation, addressing the limitations of convolutional neural networks (CNNs) in capturing long-range dependencies. The authors propose a gated axial-attention model that introduces an additional control mechanism in the self-attention module, enhancing the ability to learn global and local features. They also introduce a Local-Global (LoGo) training strategy, which operates on both the whole image and patches to improve performance. The proposed Medical Transformer (MedT) is evaluated on three medical image segmentation datasets and shown to outperform both CNNs and other transformer-based architectures. The key contributions include a gated position-sensitive axial attention mechanism, effective LoGo training methodology, and the MedT model, which demonstrates improved performance in medical image segmentation tasks.