This paper introduces the Dual Attention Network (DANet) for scene segmentation, which captures rich contextual dependencies using a self-attention mechanism. Unlike previous methods that rely on multi-scale feature fusion, DANet integrates local features with their global dependencies through two attention modules: a position attention module and a channel attention module. The position attention module models spatial interdependencies by aggregating features at each position using a weighted sum, while the channel attention module emphasizes interdependent channel maps by integrating associated features. These modules are appended to a dilated FCN to enhance feature representation and improve segmentation accuracy. The proposed method achieves state-of-the-art performance on three challenging datasets: Cityscapes, PASCAL Context, and COCO Stuff, with a Mean IoU score of 81.5% on Cityscapes. The contributions of the paper include the introduction of DANet, the design of position and channel attention modules, and the demonstration of their effectiveness through comprehensive experiments.This paper introduces the Dual Attention Network (DANet) for scene segmentation, which captures rich contextual dependencies using a self-attention mechanism. Unlike previous methods that rely on multi-scale feature fusion, DANet integrates local features with their global dependencies through two attention modules: a position attention module and a channel attention module. The position attention module models spatial interdependencies by aggregating features at each position using a weighted sum, while the channel attention module emphasizes interdependent channel maps by integrating associated features. These modules are appended to a dilated FCN to enhance feature representation and improve segmentation accuracy. The proposed method achieves state-of-the-art performance on three challenging datasets: Cityscapes, PASCAL Context, and COCO Stuff, with a Mean IoU score of 81.5% on Cityscapes. The contributions of the paper include the introduction of DANet, the design of position and channel attention modules, and the demonstration of their effectiveness through comprehensive experiments.